JP2018138159A - Image processing device, optical interference tomographic imaging apparatus, image processing method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processing device which can improve a problem related to generation of a high contrast tomographic image using statistical processing.SOLUTION: An image processing device processing a tomographic image generated by using a plurality of sets of tomographic signals acquired by performing a plurality of times of optical interference tomographic imaging by using measurement light controlled so as to scan substantially the same spots of a subject includes: an acquisition unit which acquires the plurality of sets of tomographic signals; and a generation unit which generates the tomographic image from a representative value at each pixel position of the tomographic image generated by using the plurality of sets of tomographic signals. The generation unit includes: a region identification section which identifies a target region where contrast is reduced in the tomographic image by using at least one set of tomographic signals out of the plurality of sets of tomographic signals; and an estimation section which estimates an estimation value by performing statistical processing to the plurality of sets of tomographic signals at each pixel position of the target region. The generation unit generates the tomographic image using the estimation value as the representative value in the target region.SELECTED DRAWING: Figure 6

Description

  The present invention relates to an image processing apparatus, an optical coherence tomography apparatus, an image processing method, and a program.

  In recent years, in the field of ophthalmology, an apparatus using an optical coherence tomography (OCT) (hereinafter referred to as an OCT apparatus) that can observe / measure a tomographic image of a fundus and an anterior segment non-invasively has become widespread. . The OCT apparatus is an apparatus that obtains information related to a tomogram of an eye to be inspected by using interference light obtained by irradiating the eye to be examined with low-coherent light (measurement light) and combining return light from the eye to be examined with reference light. . In the OCT apparatus, a tomographic image can be acquired by scanning low-coherent light on the fundus of the eye to be examined. OCT devices are widely used in medicine from research to clinical practice.

  OCT is roughly classified into two types, time domain OCT and Fourier domain OCT. Further, the Fourier domain OCT includes a spectral domain OCT (SD-OCT) and a swept source OCT (SS-OCT). The Fourier domain OCT can acquire information on a tomogram of an object by using a light source having a wide wavelength band, performing signal acquisition by dispersing interference light and performing Fourier transform or the like on the acquired signal. . In SD-OCT, light is spatially dispersed by a spectroscope. In contrast, in SS-OCT, light is temporally dispersed using light sources that emit light having different wavelengths.

  The tomographic images of the anterior ocular segment and fundus acquired by OCT can observe the state of these tissues only with a single tomographic image. However, it is possible to observe / measure in detail only a single tomographic image of the vitreous body, which is a tissue with a small refractive index difference that can only acquire weak signals, and the choroid that exists deep in the fundus tissue. However, it is difficult because it is buried in the noise of the device.

  In order to solve this problem, conventionally, by averaging a plurality of tomographic images acquired by imaging the same portion of a subject multiple times, the noise of a single tomographic image is suppressed, and a finer tomographic image is obtained. Techniques for generating images have been taken.

  On the other hand, recent studies have revealed that the tomographic image generated by this averaging has a low contrast in a region where the signal intensity is low in principle. According to Non-Patent Document 1, since the signal intensity of OCT has a Rice distribution, it is understood that the signal intensity is biased in a direction in which the signal intensity increases by averaging the signal intensity particularly in a region where the signal intensity is low. Yes. For this reason, the averaged tomographic image is imaged based on a signal strength higher than the actual signal strength in a region where the signal strength is low, and the contrast between the high signal strength region and the low region is low. End up.

  In view of the fact that contrast is lowered by averaging of tomographic images, Non-Patent Document 1 proposes a method for generating a tomographic image using MAP estimation (Maximum a posteriori estimate) processing which is a kind of statistical processing. Has been. In this method, compared with the averaging of tomographic images, a value closer to the actual signal is estimated as a signal intensity and a tomographic image is generated by MAP estimation, thereby obtaining a high-contrast tomographic image. According to Non-Patent Document 1, a tomographic image suitable for detailed observation / measurement is generated by this method, which has a higher contrast of the choroid in the deep fundus than the conventional averaged tomographic image. In the method described in Non-Patent Document 1, the likelihood related to the signal strength of each pixel in a plurality of tomographic images at the same location is calculated for all the pixels of the tomographic image, and the signal strength that maximizes the product of the likelihoods is calculated. Estimated as true signal strength.

Chan AC, Kurokawa K, Makita S, Miura M, Yasuno Y., Maximum a posteriori estimator for high-contrast image composition of optical coherence tomography, Opt Lett., 2016; 41 (2): 321, doi: 10.1364 / OL. 41.000321.

  However, in the method described in Non-Patent Document 1, since the above estimation is performed for all pixels of the tomographic image, the amount of calculation becomes very large, and it takes time to calculate the result. For this reason, since it takes time to determine whether or not imaging has been properly performed, for example, the subject may be left with a face on the OCT apparatus for a long time. May give. In addition, for an examiner who operates the OCT apparatus, for example, since it is slow to determine whether or not photographing has been properly performed, it takes a long time for one person to photograph, making it difficult to photograph many subjects. Problem arises.

  Further, in the MAP estimation process described in Non-Patent Document 1, it is assumed that the distribution of the signal strength of OCT acquired with respect to the true signal strength is based on the signal characteristics and noise characteristics, and the signal based on the distribution of a large number of signal acquisition results The true signal strength that maximizes the appearance probability of is estimated. For this reason, for example, a result that is not correctly measured in the signal acquisition result due to an external factor such as the eye moving during measurement and the intensity of the reflected signal from the eye changing, or the amount of noise changing due to disturbance. If it is included, it leads to fluctuation of the estimation result.

  Here, when the measured signal strength is large with respect to the change in signal strength caused by an external factor, the influence is very small. However, when the measured signal strength is small, a minute change in signal strength due to an external factor may cause a large fluctuation in the estimation result. As a result, for example, when a tomographic image of the fundus is taken, roughness may be conspicuous on the image, particularly in the vitreous body, sclera, and choroid, which are regions with low signal intensity.

  Furthermore, since the high contrast of the tomographic image by the MAP estimation process described in Non-Patent Document 1 is a statistical estimation technique, the more pixels that are used (the number of tomographic images), the more reliable the true value. Can be estimated.

  However, the subject must naturally keep his eyes open while taking a tomographic image of the fundus. Therefore, when taking a large number of tomographic images, the subject is extremely burdened. In addition, since it takes a long time to take a large number of tomographic images, the subject is burdened with time. In addition to this, if the imaging time becomes longer, there is a higher possibility that the imaging will fail, and there will be a further burden on the subject, such as the need to perform reimaging.

  Accordingly, the present invention provides an image processing apparatus, an optical coherence tomography apparatus, an image processing method, and a program that can improve at least one of the above-described problems related to generation of a high-contrast tomographic image using statistical processing. To do.

  An image processing apparatus according to an embodiment of the present invention uses a plurality of sets of tomographic signals acquired by performing optical coherence tomography multiple times using measurement light controlled to scan substantially the same part of a subject. An image processing apparatus that processes the tomographic images generated in step S10, and generates a representative value at each pixel position of the tomographic image using the acquisition unit that acquires the plurality of sets of tomographic signals and the plurality of sets of tomographic signals. And a generating unit that generates the tomographic image using a representative value at each pixel position as a pixel value, and the generating unit uses the at least one set of tomographic signals among the plurality of sets of tomographic signals to generate the tomographic image. An area specifying unit that specifies a target area where contrast is reduced in an image, and a statistical process on a plurality of tomographic signals corresponding to the pixel position among the plurality of sets of tomographic signals at each pixel position of the target area. By performing, and a estimation unit that estimates an estimated value, the generating unit includes at the target region to generate the tomographic image the estimated value as the representative value.

  An image processing apparatus according to another embodiment of the present invention is an image processing apparatus that processes a tomographic image generated using a tomographic signal acquired by optical coherence tomographic imaging of a subject. An acquisition unit for acquiring, and a generation unit for generating the tomographic image using the tomographic signal, wherein the generation unit performs statistical processing on a tomographic signal corresponding to a pixel position of the tomographic image among the tomographic signals By performing the above, an estimation unit that estimates an estimated value and a filter unit that applies a noise removal filter to a tomographic image generated using the estimated value as a pixel value are included.

  An image processing apparatus according to still another embodiment of the present invention is an image processing apparatus that processes a tomographic image generated using a tomographic signal acquired by optical coherence tomographic imaging of a subject. An acquisition unit that acquires a signal, and a generation unit that generates the tomographic image using the tomographic signal, wherein the generation unit applies to a tomographic signal corresponding to a pixel position of the tomographic image of the tomographic signal An estimator that estimates an estimated value by performing statistical processing, wherein the estimator converts to a tomographic signal corresponding to a pixel position of the tomographic signal that estimates the estimated value and a pixel position around the pixel position; The estimated value is estimated by performing statistical processing on the image, and the generation unit generates the tomographic image using the estimated value as a pixel value.

  An image processing method according to still another embodiment of the present invention is an image processing method for processing a tomographic image generated using a tomographic signal acquired by optical coherence tomographic imaging of a subject, the tomographic signal And the step of generating the tomographic image using the tomographic signal, wherein the step of generating the tomographic image uses the tomographic signal to determine a target region where contrast is reduced in the tomographic image. Including the step of identifying and estimating the estimated value by performing statistical processing on the tomographic signal of the target region of the tomographic signal, and in the step of generating the tomographic image, the target region Then, the tomographic image is generated using the estimated value as a pixel value.

  According to the present invention, there are provided an image processing apparatus, an optical coherence tomographic imaging apparatus, an image processing method, and a program capable of improving problems relating to generation of a high-contrast tomographic image using statistical processing.

1 shows a schematic configuration of an OCT apparatus according to Embodiment 1 of the present invention. 1 shows a schematic configuration of an imaging optical system according to a first embodiment. 1 shows a schematic configuration of a control unit according to a first embodiment. An example of the display screen of the display part by Example 1 is shown. An example of the display screen of the display part by Example 1 is shown. 3 shows a flowchart of tomographic image photographing processing according to the first embodiment. FIG. 6 is a diagram for explaining a tomographic image photographing process according to the first embodiment. 3 shows a flowchart of a MAP estimation process according to the first embodiment. An example of a probability density map is shown. An example of the profile of A scan data obtained by the tomographic imaging process according to the first embodiment is shown. 9 shows a flowchart of a tomographic image photographing process according to Embodiment 2. An example of the display screen of the display part by Example 2 is shown. 9 shows a flowchart of tomographic image photographing processing according to Embodiment 3. It is a figure for demonstrating the area | region identification process by the modification of Example 1 thru | or 3. 10 shows a schematic configuration of a control unit according to a fourth embodiment. An example of the display screen of the display part which concerns on Example 4 is shown. 9 is a flowchart of tomographic image capturing processing according to Embodiment 4. 10 shows a flowchart of a MAP estimation process according to Embodiment 4. It is a figure for demonstrating the noise reduction which concerns on Example 4. FIG. 10 shows a schematic configuration of a control unit according to a fifth embodiment. 10 is a flowchart of tomographic imaging processing according to Embodiment 5. FIG. 10 is a diagram for explaining region specifying processing according to a fifth embodiment. 10 is a flowchart of tomographic image capturing processing according to a modification of the fifth embodiment. FIG. 10 is a diagram for explaining region specifying processing according to a sixth embodiment. 10 shows a schematic configuration of a control unit according to a seventh embodiment. 10 is a flowchart of tomographic image capturing processing according to Embodiment 7. 10 is a flowchart of range MAP estimation processing according to Embodiment 7. It is a figure for demonstrating noise reduction. It is a figure for demonstrating the weighting which concerns on the modification of Example 7. FIG. 9 shows a schematic configuration of a control unit according to an eighth embodiment. 10 is a flowchart of tomographic imaging processing according to Embodiment 8. FIG. 10 is a diagram for explaining region specifying processing according to an eighth embodiment.

  Hereinafter, exemplary embodiments and examples for carrying out the present invention will be described in detail with reference to the drawings. However, dimensions, materials, shapes, and relative positions of components described in the following embodiments and examples are arbitrary, and can be changed according to the configuration of the apparatus to which the present invention is applied or various conditions. . Also, in the drawings, the same reference numerals are used between the drawings to indicate the same or functionally similar elements.

(Embodiment 1)
As described above, in the MAP estimation process described in Non-Patent Document 1, since the above estimation is performed for all pixels of the tomographic image, the amount of calculation becomes very large, and it takes time to calculate the result. For this reason, there are cases where physical burdens are imposed on the subject and it is difficult to photograph a large number of subjects. Therefore, Embodiment 1 of the present invention provides an optical coherence tomographic imaging apparatus that can generate a tomographic image having a higher contrast than an averaged image in a shorter time.

  Hereinafter, before describing an example of the present embodiment, the present embodiment will be schematically described. In this embodiment, in order to generate a high-contrast tomographic image in a short time, the amount of calculation that requires a great amount of time is reduced by limiting the pixels for which MAP estimation is performed. The reason why a high-contrast tomographic image can be generated in a shorter time compared to the method described in Non-Patent Document 1 according to the present embodiment will be described below.

  In the Rice distribution when the signal intensity is low, bias is generated in the direction in which the value increases in the signal intensity when averaging is performed due to the convex asymmetry with respect to the positive direction of the signal intensity. Therefore, as described above, when a tomographic image is generated by averaging a plurality of tomographic images, a tomographic image is generated based on a signal strength higher than the actual signal strength in an area where the signal strength is low. For such a region, the method disclosed in Non-Patent Document 1 can calculate a value closer to the actual signal strength by MAP estimation. Here, the signal intensity mainly refers to a luminance value which is a pixel value of a pixel, but may be a signal value corresponding to the pixel. Therefore, the signal strength includes, for example, the strength of the signal after the Fourier transform of the interference signal and the signal obtained by performing some signal processing on the signal.

  On the other hand, conventionally, it has been known that in the Rice distribution, if the actual signal value (signal intensity) is sufficiently large with respect to noise, the distribution approaches a Gaussian distribution. Therefore, it is understood that when the signal intensity is high, there is no difference between the results of the OCT tomographic image averaging and the MAP estimation.

  Here, in order to make the conventional averaged OCT tomographic image a higher-contrast image, by performing MAP estimation on pixels with low actual signal intensity, compared with the corresponding pixels of the averaged image, It suffices if the pixel value can be set to a low value closer to the actual signal intensity. On the other hand, for pixels with high signal strength, there is no difference between the result of MAP estimation and the averaged result, so there is no need to perform MAP estimation.

  As described above, when MAP estimation is performed, the amount of calculation is significantly larger than when averaging is performed. Therefore, averaging processing is performed on pixels with high signal intensity, and MAP estimation processing is performed only on pixels with low signal intensity, so that a high-contrast tomography equivalent to the case where MAP estimation is performed on all pixels of a tomographic image. An image can be generated in a shorter time.

  In the present embodiment, the basis of the judgment for dividing the processing content for each pixel by MAP estimation and averaging is the signal intensity at each pixel. For example, it is conceivable that a signal intensity histogram is calculated using the averaged tomographic image, and the lower 10th percentile of the histogram is used as a threshold, and MAP estimation is performed when the signal intensity is equal to or lower than the threshold. As another method, a threshold setting method using the shape of a histogram is also conceivable.

  As another method, a determination method using a segmentation function for identifying the layer structure of the fundus tissue provided in the conventional OCT is also conceivable. For example, from the structure of the eye, the signal intensity is low in the vitreous body where the signal intensity is low because the difference in refractive index is small, or the choroid that is difficult for OCT measurement light to reach due to the deep fundus. Known as. Therefore, a method of performing segmentation processing on a tomographic image, extracting a vitreous body or choroid region, performing MAP estimation processing on the extracted region, and performing averaging processing on other regions may be considered.

  By performing such processing, it is possible to specify a region considered to have low signal strength. In the present embodiment, a high-contrast tomographic image can be generated in a short time by performing MAP estimation processing that requires a large amount of calculation only for the specified region.

Example 1
Hereinafter, an optical coherence tomography apparatus (OCT apparatus) according to Example 1 of the present embodiment will be described with reference to FIGS. FIG. 1 shows a schematic configuration of an OCT apparatus 1 according to the present embodiment. The OCT apparatus 1 includes an imaging optical system 100, a control unit 200 (image processing apparatus), and a display unit 300.

  The imaging optical system 100 irradiates the subject eye, which is a subject, with measurement light, detects interference light between the return light from the subject eye and the reference light, and generates an interference signal. The control unit 200 is connected to the imaging optical system 100 and the display unit 300 and controls them. The control unit 200 can acquire the interference signal generated by the imaging optical system 100 and generate a tomographic image of the eye to be examined. The display unit 300 can display various images, information on the eye to be examined, and the like sent from the control unit 200.

  Here, the control unit 200 may be configured using a general-purpose computer, or may be configured as a dedicated computer for the imaging optical system 100. The display unit 300 can be configured using any display. In the present embodiment, the control unit 200 and the display unit 300 are configured separately from the imaging optical system 100, but all or part of them may be configured integrally.

(Configuration of imaging optical system)
Next, the configuration of the imaging optical system 100 according to the present embodiment will be described with reference to FIG. FIG. 2 shows a schematic configuration of the imaging optical system 100.

  In the imaging optical system 100, the objective lens 101 is installed facing the eye 140 to be examined. On the optical axis of the objective lens 101, a dichroic mirror 102 and a dichroic mirror 103 are provided. The optical path from the objective lens 101 is light passing through the optical path by the dichroic mirrors 102 and 103 to the optical path L1 of the OCT optical system, the optical path L2 of the internal fixation lamp and the fundus observation system, and the optical path L3 of the anterior ocular segment observation system. Is branched for each wavelength band. In the present embodiment, the optical path L1 of the OCT optical system and the optical path L2 of the internal fixation lamp and the fundus oculi observation system are arranged in the transmission direction of the dichroic mirror 102, and the optical path L3 of the anterior ocular segment observation system is arranged in the reflection direction. An optical path L2 of the internal fixation lamp and the fundus oculi observation system is arranged in the transmission direction of the dichroic mirror 103, and an optical path L1 of the OCT optical system is arranged in the reflection direction. However, the arrangement of these optical paths is not limited to this arrangement, and the respective optical paths may be arranged so as to be opposite to the transmission direction and the reflection direction of the dichroic mirrors 102 and 103, respectively.

  In the optical path L2, lenses 104 and 105, a dichroic mirror 134, an internal fixation lamp 106, and a CCD 111 for fundus observation are provided. In addition, the component arrange | positioned in the optical path of a fundus observation system constitutes a fundus observation optical system. The lens 104 is a focusing lens, and is driven in the optical axis direction indicated by an arrow in the drawing by a motor (not shown) controlled by the control unit 200 for adjusting the focusing of light passing through the optical path L2. The optical path L2 is branched by the dichroic mirror 134 into an optical path to the internal fixation lamp 106 and an optical path to the CCD 111. In this embodiment, the internal fixation lamp 106 is disposed in the transmission direction of the dichroic mirror 134, and the CCD 111 is disposed in the reflection direction. The CCD 111 may be arranged in the transmission direction of the dichroic mirror 134 and the internal fixation lamp 106 may be provided in the reflection direction.

  The CCD 111 has sensitivity at the wavelength of illumination light for fundus observation (not shown), specifically around 780 nm. The internal fixation lamp 106 is used to generate visible light and promote fixation of the subject.

  The return light emitted from the fundus observation light source and reflected by the eye 140 is transmitted through the objective lens 101 and the dichroic mirrors 102 and 103 and enters the optical path L2. The return light incident on the optical path L 2 passes through the lenses 104 and 105, is reflected by the dichroic mirror 134, and is guided to the CCD 111. The CCD 111 detects the return light from the incident eye 140 and generates a signal corresponding to the return light. The control unit 200 can obtain a front image of the fundus Er of the eye to be examined 140 based on the signal generated by the CCD 111.

  The light emitted from the internal fixation lamp 106 passes through the dichroic mirror 134, the lenses 105 and 104, the dichroic mirrors 103 and 102, and the objective lens 101 and enters the eye 140 to be examined. The internal fixation lamp 106 can provide light of an arbitrary shape as a fixation target at an arbitrary position on the eye 140, and can promote fixation of the subject.

  Note that the configuration of the fundus oculi observation optical system is not limited to the above-described configuration. For example, the fundus observation optical system may have a configuration of an SLO (Scanning Laser Ophthalmoscope) that scans the eye with illumination light. In this case, by flashing the internal fixation lamp 106 in accordance with the movement of the scanning means of the SLO optical system, it is possible to provide light of an arbitrary shape as a fixation target at an arbitrary position on the eye 140 to be examined. It is possible to promote fixation of the subject.

  Next, the optical path L3 of the anterior segment observation system will be described. In the optical path L3 of the anterior ocular segment observation system, lenses 107 and 109, a split prism 108, and an anterior ocular segment observation CCD 110 that detects infrared light are provided. These components arranged in the optical path L3 of the anterior ocular segment observation system constitute an anterior ocular segment observation optical system (imaging means).

  In the optical path L <b> 3, an anterior ocular segment observation light having a wavelength near 970 nm is irradiated to the anterior ocular segment of the eye 140 to be examined from a light source (not shown). Reflected light from the anterior segment of the eye 140 to be examined enters the split prism 108 via the objective lens 101, the dichroic mirror 102, and the lens 107. The split prism 108 is disposed at a position conjugate with the pupil of the eye 140 to be examined. The light emitted from the split prism 108 enters the CCD 110 via the lens 109.

  The CCD 110 detects light having a wavelength near 970 nm, detects reflected light from the anterior segment, and generates a signal corresponding to the reflected light from the anterior segment. The control unit 200 can generate an image of the anterior segment of the eye to be examined 140 based on the signal generated by the CCD 110. At this time, the control unit 200 detects the reflected light that has passed through the split prism 108 by the CCD 110, thereby determining the distance in the Z direction (depth direction) of the imaging optical system 100 from the split image of the anterior eye part to the eye 140 to be examined. Can be detected.

  Next, the optical path L1 will be described. The optical path L1 constitutes the optical path of the OCT optical system as described above, and is used to acquire an interference signal for forming a tomographic image of the eye 140 to be examined. In the optical path L1, an X scanner 131, a Y scanner 132, and lenses 112 and 113 are arranged.

  The X scanner 131 and the Y scanner 132 constitute scanning means for scanning the measurement light on the fundus Er of the eye 140 to be examined. The X scanner 131 and the Y scanner 132 are driven by a galvano motor (not shown) controlled by the control unit 200. The X scanner 131 is used for scanning the measurement light in the X direction, and the Y scanner 132 is used for scanning the measurement light in the Y direction. Note that the X scanner 131 and the Y scanner 132 can be configured using an arbitrary deflection mirror such as a galvanometer mirror. In the present embodiment, the scanning unit is configured by the X scanner 131 and the Y scanner 132, but the configuration of the scanning unit is not limited to this. The scanning unit may be configured by a deflecting mirror that can deflect light in a two-dimensional direction with a single sheet such as a MEMS mirror.

  The lens 112 is a focusing lens used for adjusting the focus of measurement light emitted from the optical fiber 115 of the OCT measurement optical system with respect to the fundus Er of the eye 140 to be examined. The lens 112 is driven in the optical axis direction of measurement light indicated by an arrow in the figure by a motor (not shown) controlled by the control unit 200. By this focusing adjustment, the return light from the fundus Er is simultaneously focused on the tip of the optical fiber 115 and made incident. Note that the optical fiber 115, the optical members disposed in the optical path L1, the dichroic mirrors 102 and 103, the objective lens 101, and the like constitute an OCT measurement optical system through which measurement light propagates in the OCT optical system.

  The optical fiber 115 is connected to the optical coupler 119. The optical coupler 119 is connected to the optical fiber 115 of the OCT measurement optical system, the optical fiber 116 connected to the light source 114, the optical fiber 117 of the OCT reference optical system, and the optical fiber 118 connected to the spectroscope 123. The optical coupler 119 functions as a splitter that splits light from the light source 114 into measurement light and reference light, and an interference unit that causes interference between the return light of the measurement light from the eye 140 and the reference light to generate interference light. To do.

  The light source 114 is an SLD (Super Luminescent Diode) which is a typical low coherent light source. In this embodiment, a light source 114 having a center wavelength of emitted light of 855 nm and a wavelength bandwidth of about 100 nm is used. Note that the configuration of the light source 114 is not limited to this, and an arbitrary light source can be used according to a desired configuration.

  The light emitted from the light source 114 is split through the optical fiber 116 through the optical coupler 119 into measurement light propagating through the OCT measurement optical system such as the optical fiber 115 and reference light propagating through the OCT reference optical system such as the optical fiber 117. The The measurement light is irradiated onto the fundus Er of the eye 140 to be examined through the optical path L1 of the OCT optical system described above, and reaches the optical coupler 119 through the same optical path as return light due to reflection and scattering by the retina.

  On the other hand, the reference light reaches the reference mirror 122 and is reflected through the optical fiber 117, the lens 120, and the dispersion compensation glass 121 inserted to match the dispersion of the measurement light and the reference light. Then, it returns on the same optical path and reaches the optical coupler 119. Here, the optical fiber 117, the lens 120, the dispersion compensation glass 121, and the reference mirror 122 constitute an OCT reference optical system.

  The return light and the reference light of the measurement light from the eye 140 to be examined are combined by an optical coupler 119 to become interference light. When the optical path length of the measurement light and the optical path length of the reference light are substantially equal, the return light of the measurement light and the reference light interfere with each other and become interference light. The reference mirror 122 is held by an unillustrated motor and drive mechanism controlled by the control unit 200 so as to be adjustable in the optical axis direction of the reference light indicated by an arrow in the figure, and changes depending on the measured part of the eye 140 to be examined. It is possible to match the optical path length of the reference light with the optical path length of the light. The interference light is guided to the spectroscope 123 via the optical fiber 118.

  The spectroscope 123 (light detection unit) includes lenses 124 and 126, a diffraction grating 125, and a line sensor 127. The interference light emitted from the optical fiber 118 becomes substantially parallel light through the lens 124, and then is split by the diffraction grating 125 and imaged on the line sensor 127 by the lens 126. The line sensor 127 is shown as an example of a light receiving element that receives interference light, generates an output signal corresponding to the interference light, and outputs the output signal. Based on the signal generated by the line sensor 127, the control unit 200 can acquire information related to the tomographic image of the fundus Er of the eye to be examined 140 and generate a tomographic image of the fundus Er.

(Configuration of control unit 200)
The configuration of the control unit 200 will be described with reference to FIG. FIG. 3 is a block diagram illustrating a schematic configuration of the control unit 200. The control unit 200 includes an image generation unit 210, an acquisition unit 220, a drive control unit 230, a storage unit 240, and a display control unit 250.

  The acquisition unit 220 acquires various signals from the CCDs 110 and 111 and the line sensor 127 of the imaging optical system 100. The acquisition unit 220 can also acquire, from the tomographic image generation unit 213, a signal after Fourier transform generated based on the interference signal, a signal obtained by performing some signal processing on the signal, and the like.

  The image generation unit 210 is provided with an anterior eye image generation unit 211, a fundus image generation unit 212, a tomographic image generation unit 213, a region specifying unit 214, a calculation unit 215, and an estimation unit 216.

  The anterior eye image generation unit 211 generates an anterior eye image of the eye to be examined 140 based on the signal from the CCD 110 acquired by the acquisition unit 220. The fundus image generation unit 212 generates a fundus image of the eye 140 to be examined based on the signal from the CCD 111 acquired by the acquisition unit 220.

  The tomographic image generation unit 213 generates a tomographic image of the eye to be examined 140 based on the interference signal from the line sensor 127 acquired by the acquisition unit 220. More specifically, the tomographic image generation unit 213 performs Fourier transform on the interference signal obtained from the line sensor 127 by the series of processes described above, and converts the signal after Fourier transform into luminance or density information. Thereby, the tomographic image generation unit 213 acquires a tomographic image in the depth direction (Z direction) at a certain point of the fundus Er of the eye 140 to be examined. Such a scanning method is called A scan, and the obtained tomographic image is called A scan image.

  A plurality of A-scan images can be acquired by repeatedly performing such A-scan while scanning the measurement light in the predetermined transverse direction of the fundus Er with the X scanner 131 and the Y scanner 132. For example, if the measurement light is scanned in the X direction by the X scanner 131, a tomographic image on the XZ plane is obtained, and if scanned in the Y direction, a tomographic image on the YZ plane is obtained. A method of scanning the fundus Er of the eye 140 to be examined in a predetermined transverse direction is called a B scan, and the obtained tomographic image is called a B scan image.

  In addition, the tomographic image generation unit 213 generates a tomographic image with reduced noise using the calculation value calculated by the calculation unit 215. Furthermore, the tomographic image generation unit 213 generates a high-contrast tomographic image using the calculated value calculated by the calculating unit 215 and the estimated value estimated by the estimating unit 216.

  Based on the tomographic image generated by the tomographic image generation unit 213, the region specifying unit 214 specifies a target region having a low signal intensity (region where contrast is lowered). Note that the region specifying unit 214 may specify the target region based on the signal after Fourier transform acquired by the acquiring unit 220 or the like.

  The calculation unit 215 calculates a calculation value at each pixel position based on the plurality of tomographic images generated by the tomographic image generation unit 213. In the present embodiment, the calculation unit 215 calculates an addition average value (arithmetic average value) of luminance values at each pixel position of a plurality of tomographic images as a calculation value. The calculated value is not limited to the addition average value, and may be a median value, a mode value, a maximum value, or the like, for example. Note that the calculation unit 215 may calculate the calculation value at each pixel position based on the signal after Fourier transform acquired by the acquisition unit 220 or the like.

  The estimation unit 216 performs MAP estimation processing on the plurality of tomographic images generated by the tomographic image generation unit 213, and estimates the estimated value of the true value of the interference signal. Note that the estimation unit 216 may estimate the estimated value at each pixel position based on the signal after Fourier transform acquired by the acquisition unit 220 or the like.

  The drive control unit 230 controls driving of each component of the imaging optical system 100 such as the light source 114, the X scanner 131, and the Y scanner 132. The storage unit 240 stores various images generated by the image generation unit 210, information on the input subject, programs that constitute the control unit 200, and the like. The display control unit 250 controls the display unit 300 and causes the display unit 300 to display various images stored in the storage unit 240 and information on the subject.

  Each component of the control unit 200 can be configured by a module executed by the CPU or MPU of the control unit 200. In addition, each component of the control unit 200 may be configured by a circuit that implements a specific function such as an ASIC. The storage unit 240 can be configured using any storage device / storage medium such as a memory or an optical disk.

(Tomographic imaging process)
Hereinafter, with reference to FIGS. 4A to 9, a tomographic image capturing process according to the present embodiment will be described. FIG. 4A shows an example of a control / image display GUI preview screen 400 displayed on the display unit 300. FIG. 4B shows an example of a control / image display GUI report screen 406 displayed on the display unit 300. A preview screen 400 illustrated in FIG. 4A is a screen displayed for performing an instruction to start imaging, alignment adjustment of the imaging optical system 100, adjustment of the position of a part to be imaged, and the like. On the other hand, a report screen 406 shown in FIG. 4B is a screen for displaying high-quality tomographic images 407 and 408 generated by imaging. In this embodiment, a function of generating a high-contrast tomographic image by pressing a high-contrast mode button 405 on the preview screen 400 shown in FIG. 4A and turning on the high-contrast mode using an input device (not shown). Can be used.

  In this embodiment, after preparing for photographing on the preview screen 400 shown in FIG. 4A, the fundus Er of the subject is photographed using the OCT apparatus 1. Specifically, first, the face of the subject is placed on the face receiving portion of the imaging optical system 100, and the imaging optical system 100 is aligned with the eye 140 so that the measurement light enters the pupil of the eye 140 to be examined. In this embodiment, the examiner moves the imaging optical system 100 in the XYZ directions using a driving stage (not shown) while viewing the anterior eye image 401 on the preview screen 400 and the fundus image 402 displayed on the fundus preview. Then, the imaging optical system 100 is aligned. The control unit 200 may analyze the anterior eye image 401 and the like, and control the driving of the imaging optical system 100 based on the analysis result to align the imaging optical system 100 with the eye 140 to be examined.

  Next, a scan mode in which the substantially same portion displayed in the scan mode 403 is imaged a plurality of times is selected, and imaging is started by pressing the imaging button 404 or the imaging button attached to the imaging optical system 100.

  FIG. 5 shows a flowchart of tomographic image capturing processing according to the present embodiment. When shooting is started, the process proceeds to step S501. In step S501, the tomographic image generation unit 213 acquires background data as a noise signal in the OCT optical system. The background data is acquired, for example, by blocking the return light of the measurement light and acquiring the optical signal of only the reference light. Note that the tomographic image generation unit 213 may acquire noise information such as background data that is measured in advance for each apparatus and stored in the storage unit 240.

  In step S <b> 502, the imaging optical system 100 images N times of substantially the same portion (scanning line) of the fundus Er of the eye 140 to be examined. Specifically, the drive control unit 230 controls each component of the imaging optical system 100 to perform N-time optical coherence tomographic imaging of substantially the same portion of the eye 140 to be examined. Thereby, the acquisition unit 220 acquires N sets of interference signals obtained by N B scans. In this embodiment, N = 50 and 50 times of optical coherence tomographic imaging are performed, but the number of imaging is not limited to this. N, which is the number of times of imaging, may be an arbitrary number of 2 or more so that a desired number of tomographic images for performing MAP estimation processing and addition averaging processing of tomographic images can be obtained. In the following, a set of interference signals corresponding to one B scan is referred to as one set of interference signals. It should be noted that “substantially the same place” includes not only the completely same place but also a place slightly deviated from the same place.

  In step S503, the tomographic image generation unit 213 generates one tomographic image based on each set of interference signals acquired by the acquisition unit 220, and generates a total of N tomographic images. The tomographic image generation unit 213 performs a Fourier transform on the set of interference signals acquired by the acquisition unit 220, converts the signal after the Fourier transform into luminance or density information, and generates a single tomographic image. The tomographic image generation process may be performed by any known processing method.

  Next, in step S504, the tomographic image generation unit 213 aligns the acquired N tomographic images of the fundus oculi N at substantially the same location by using features in the tomographic image such as the fundus shape.

  Specifically, first, the tomographic image generation unit 213 selects any one of N pieces of tomographic images as a template. For example, the tomographic image generation unit 213 can select the first generated tomographic image as the tomographic image to be selected as the template. Further, the tomographic image generation unit 213 calculates correlations in all combinations among the N pieces of tomographic images, calculates the sum of correlation coefficients for each frame, and selects the tomographic image having the maximum sum as a template. Good.

  Next, the tomographic image generation unit 213 collates with the template for each tomographic image and obtains a positional deviation amount (δX, δY, δθ) for each tomographic image. Here, δX represents a displacement amount in the X direction, δY represents a displacement amount in the Y direction, and δθ represents a displacement amount in the rotation direction. Specifically, the tomographic image generation unit 213 calculates Normalized Cross-Correlation (NCC), which is an index representing the degree of similarity with each frame while changing the position and angle of the template. The tomographic image generation unit 213 obtains the position difference between the position of the tomographic image to be verified and the template when the calculated NCC value is maximum. Note that the index indicating the similarity between images may be a scale indicating the similarity between the features of the image of the frame to be matched and the template image, and can be variously changed to any index indicating such a scale. It is.

  The tomographic image generation unit 213 aligns the tomographic images by applying position correction to N−1 tomographic images other than the template in accordance with the obtained positional deviation amounts (δX, δY, δθ). In some cases, such as when the subject's eyes have moved significantly, an image whose position is not aligned may be acquired when performing image alignment using the fundus shape. In such a case, the processing of the image whose position is not properly matched is stopped, and the processing is advanced by reducing the number of tomographic images used for the subsequent processing. As an alternative, the number of tomographic images for which processing has been stopped may be additionally acquired, and the additional tomographic images may be aligned. As a result of the alignment of the N tomographic images, when the coordinates of the pixel in each image are the same, the position of the fundus Er displayed on the pixel is also the same position. When the alignment of the tomographic image is completed, the process proceeds to step S505.

  In step S505, the calculation unit 215 calculates a calculation value of N pixel values for each pixel with respect to the aligned N tomographic images, and calculates the calculation value calculated by the tomographic image generation unit 213 for each pixel. An averaged tomographic image (average tomographic image) is generated using the representative value. The storage unit 240 stores the generated average tomographic image. In the present embodiment, the calculation unit 215 calculates the calculated value by averaging the luminance values at the pixel positions in the N tomographic images corresponding to the pixel positions of the generated average tomographic image. Then, the tomographic image generation unit 213 generates an average tomographic image using the calculated calculation value as a luminance value (pixel value) at each pixel position. Note that, as described above, the calculated value is not limited to an arithmetic average value (arithmetic average value), and may be a median value, a mode value, a maximum value, or the like.

  FIG. 6A shows an example of the generated average tomographic image. In the generated average tomographic image, as described above, the region where the signal intensity is low is expressed in a state where the contrast is low. Therefore, in the average tomographic image shown in FIG. 6A, the vitreous body 601 is displayed with a low contrast. When the average tomographic image is generated in step S505, the process proceeds to step S506.

  In step S506, the tomographic image generation unit 213 determines whether or not the high contrast mode is turned on. If the tomographic image generation unit 213 determines that the high contrast mode is not turned on, the display control unit 250 displays the average tomographic image stored in the storage unit 240 in step S505 on the display unit 300 in step S511. The imaging process ends. On the other hand, if the tomographic image generation unit 213 determines in step S506 that the high contrast mode is ON, the process proceeds to step S507.

  In step S507, the region specifying unit 214 performs a segmentation process on the average tomographic image, and specifies a vitreous region that is a region having a low signal intensity as a target region. Specifically, the region specifying unit 214 generates an image by applying a median filter and a Sobel filter to the average tomographic image (hereinafter, also referred to as a median image and a Sobel image, respectively). Next, the area specifying unit 214 generates a profile for each data corresponding to the A scan from the generated median image and Sobel image. The generated profile is a luminance value profile for a median image and a gradient profile for a Sobel image. Then, the region specifying unit 214 detects a peak in the profile generated from the Sobel image. The region specifying unit 214 extracts the boundary of each region of the retinal layer by referring to the profile of the median image before and after the detected peak and between the peaks.

  In the present embodiment, the region specifying unit 214 extracts an internal limiting membrane (ILM) by segmentation processing. The region specifying unit 214 determines that the opposite side (pupil side) of the retina at the ILM boundary is the vitreous body, and the signal intensity is determined from the region on the pupil side from the ILM boundary (region 602 indicated by hatching in FIG. 6B). It is identified as the vitreous region known as the low region. Thereby, the region specifying unit 214 specifies the vitreous region as a target region for which MAP estimation is performed for the pixel value.

  Note that the region specifying unit 214 may perform segmentation processing on one of the N tomographic images, specify a region with low signal intensity, and set the region as a target region. The segmentation process may be performed using any known method other than the above method. Furthermore, in this embodiment, the vitreous body region is the target region, but an arbitrary region such as a choroid region known as a low signal intensity region may be specified as the target region. In this case, the region specifying unit 214 can specify a layer such as a choroid by a segmentation process, and can set the layer as a target region. Furthermore, a plurality of regions such as a vitreous region and a choroid region may be set as target regions.

  Next, in step S508, the estimation unit 216 performs a MAP estimation process for the identified target region, and estimates an estimated value at each pixel position of the target region. Hereinafter, the MAP estimation according to the present embodiment will be described with reference to FIG. FIG. 7 shows a flowchart of the MAP estimation process according to this embodiment.

When the MAP estimation process is started, in step S701, the estimation unit 216 calculates the noise variance σ ² based on the noise signal acquired in step S501.

ここで、式１は、測定される信号強度の確率密度関数であり、ａ_ｎは測定した画素における信号強度、νは真の信号強度、σ^２はノイズの分散、Ｉ_０は第１種変形ベッセル関数である。なお、本実施例では測定される信号強度として、断層画像における画素の画素値（輝度値）を用いている。ここで、ｎは処理対象の信号強度のインデックスを示し、１つ目の信号強度であればｎ＝１となる。 Then, the estimation unit 216, in step S702, by utilizing the noise variance sigma ² calculated with the following equation 1 to calculate the probability density of the measured signal strength a for true signal strength [nu, true signal strength [nu Generate a probability density map of the measured signal strength a for.

Here, Equation 1 is the probability density function of the signal intensities measured, the signal intensity at the pixels measured in a _n, [nu true signal strength, sigma ² is the variance of the noise, I ₀ is the first type modification Bessel function. In this embodiment, the pixel value (luminance value) of the pixel in the tomographic image is used as the measured signal intensity. Here, n represents an index of the signal strength to be processed, and if the first signal strength, n = 1.

FIG. 8 shows an example of a probability density map of the measured signal strength A with respect to the true signal strength ν. Here, the signal strength A indicates the value of the measured signal strength a, and A ₁ , A _{2, and the} like indicate the signal strength for each step size, not the signal strength according to the index of the signal strength to be measured. That is, for example, A ₁ and A ₂ do not indicate the signal intensity of the first and second measurement objects like a ₁ and a ₂ , and A ₂ adds a value corresponding to the step size to A _1. Value. Further, the step size (width of ν _{1 to} ν ₂ , width of A _{1 to} A _2, etc.) and range (range of ν _{1 to} ν _max , A _{1 to} A _max ) of the parameter of ν and A in the probability density map (Range) is set in advance so as to have sufficient gradation and dynamic range for the tomographic image to be acquired.

The estimation unit 216 substitutes ν and A that are separated by the set step size and the calculated noise variance σ ² into Equation 1, and generates a probability density map of the signal intensity according to the substituted ν and A. Here, the signal intensity A is substituted as a _n of formula 1. In this embodiment, the estimation unit 216 obtains the probability density by sequentially changing the value of A with respect to the value of ν, and arranges the probability density in the direction of the solid line arrow in the probability density map shown in FIG. Generate a map. The estimation unit 216 obtains the probability density by sequentially changing the value of ν with respect to the value of A, and generates the probability density map by arranging the probability density in the direction of the dashed arrow in the probability density map shown in FIG. May be.

  Next, in step S703, the estimation unit 216 sets an initial position of the pixel position (coordinates (x, z)) when starting the MAP estimation. Here, coordinates (x, z) indicate coordinates in a high-contrast tomographic image (high-contrast tomographic image) to be generated, x indicates a position in the B-scan direction in the tomographic image, and z indicates a depth in the tomographic image. Indicates the position of the direction. Each coordinate in the high-contrast tomographic image to be generated corresponds to each coordinate in the average tomographic image and the N tomographic images that have been aligned. As described above, since the estimation unit 216 performs MAP estimation for the pixels in the target region specified by the region specifying unit 214, the initial position may be any position as long as it is a position in the target region. For example, the estimation unit 216 sets the coordinates having the smallest values of x and z in the target region as the initial position.

Next, in step S704~ step S706, the estimating unit 216, the probability density as a parameter the true signal strength ν for each signal strength _{a n} in the coordinates (x, z) of 1~N th tomographic image Are extracted from the probability density map. Thereafter, the estimation unit 216 calculates a probability density function using the true signal intensity ν as a parameter from the extracted probability density. Here, the probability density function using the true signal intensity ν as a parameter corresponds to a row in the example of the probability density map shown in FIG. For example, the probability density function as a parameter the true signal strength [nu, in the example shown in FIG. 8, when the value of the signal strength a _n is A ₁ is the probability for the value [nu ₃ of the signal intensity [nu = 0.2 refers to a function that leads to probability = 0.3 for ν ₄ . Note that nmax in the figure corresponds to N.

In the present embodiment, the estimation unit 216 follows the direction of the dashed arrow in the probability density map shown in FIG. 8 with respect to the signal intensity an at the coordinates (x, z) of the _nth tomographic image to be processed. To extract the probability density. When there is a tomographic image for which processing is determined to be canceled at the time of registration, the estimation unit 216 extracts probability density for the number of tomographic images excluding the number of tomographic images for which processing is determined to be stopped. To do. When estimating unit 216 extracts the probability density of the true signal intensity [nu to the signal strength a _n in N sheets of the coordinates (x, z), and calculated the probability density function as a parameter the true signal strength [nu, the process The process proceeds to step S707.

In step S707, the estimation unit 216 calculates the product of the probability density function of the true signal strength ν with respect to the N measured signal strengths from the signal strengths a ₁ to a _N , and the calculated product is the likelihood function. And Here, the product of the probability density function will be described with reference to FIG. For example, when the measured signal strength a ₁ is signal strength A ₁ and the signal strength a ₂ is signal strength A ₃ , referring to the probability density map, the true signal strength ν for the signal strengths a ₁ and a ₂ is The probability of ν ₁ is 0.1 and 0.2, respectively. Therefore, in this case, the probability that the true signal strength ν is ν ₁ is obtained as 0.1 × 0.2, which is the product of these probabilities. Similarly, ₂ true signal strength [nu is _[nu, [nu _3, ..., probability and a [nu _max are obtained respectively. As described above, a product of the probabilities of the true signal strength ν in the probability density functions of the measured signal strengths a ₁ to a _N is called a product of the probability density functions. Therefore, the product of the probability density function, ₁ true signal strength [nu is [nu from the measured signal strength a ₁ for a _N, ν _{2, ...,} and each of the product of the probabilities is [nu _max.

  In step S708, the true signal strength ν having the maximum probability in the posterior probability function (representative probability density function) that is the product of the likelihood function calculated by the estimation unit 216 and the prior probability is determined. Here, considering a uniform prior probability that a signal is always acquired as the prior probability, and assuming the prior probability as 1, the likelihood function becomes the posterior probability function as it is. The estimation unit 216 determines a parameter (true signal strength ν) having the maximum probability in the posterior probability function.

More simply, in this embodiment, the estimation unit 216 extracts a probability density for each true signal intensity ν with respect to the measured signal intensity a _{1 to} a _N in steps S704 to S706. Thereafter, in step S707, a product of probability density functions is calculated by multiplying each signal intensity by the probability density for each extracted true signal intensity ν. Then, the true signal strength ν that maximizes the probability is determined in the product of the calculated probability density functions. In this way, the determined true signal strength ν becomes an estimated value estimated by the estimating unit 216.

  In step S709, the estimation unit 216 causes the storage unit 240 to store the true signal intensity ν determined in step S708 as an estimated value of coordinates (x, z).

  Thereafter, in step S710, the estimation unit 216 determines whether or not estimated values have been estimated for all pixel positions in the target region. If the estimation unit 216 determines that the estimated values are not estimated for all pixel positions in the target region, the process proceeds to step S711. In step S711, the estimation unit 216 moves the pixel position (coordinates (x, z)) for estimating the estimated value within the target area, and then returns the process to step S704.

  If it is determined in step S710 that the estimation unit 216 has estimated the estimated values for all pixel positions in the target region, the process proceeds to step S509.

  In step S509, the tomographic image generation unit 213 overwrites the estimated value estimated by the MAP estimation processing by the estimating unit 216 as a pixel value at each pixel position of the target region in the average tomographic image. Thereby, the tomographic image generation unit 213 can generate a high-contrast tomographic image having a pixel value based on the estimated value obtained by the MAP estimation in an area where the signal intensity is low. The storage unit 240 stores the generated high contrast tomographic image.

  FIG. 6C shows an example of the generated high contrast tomographic image. In the high contrast tomographic image shown in FIG. 6C, the vitreous body 603 is shown in a high contrast state.

  FIG. 9 shows an example of a profile of signal intensity (luminance value) corresponding to the A scan in the average tomographic image generated in step S505 and the high-contrast tomographic image generated in step S509. In FIG. 9, the profile relating to the signal intensity of the average tomographic image is indicated by a solid line, and the profile relating to the signal intensity of the high contrast image using the MAP estimation process is indicated by a dotted line.

  Referring to FIG. 9, in the average tomographic image, the signal intensity is high and the contrast is low in the vitreous region where the signal intensity is originally low. In contrast, a high contrast image using the MAP estimation process uses an estimated value estimated as a true signal intensity in a vitreous region which originally has a low signal intensity, and thus a high contrast is maintained. I understand.

  In step S510, the display control unit 250 displays the high-contrast tomographic image stored in the storage unit 240 on the display unit 300, and the imaging process ends.

  As described above, the control unit 200 performs the MAP estimation process only on the vitreous region having a low signal intensity, and performs the averaging process on other regions such as the retina having a high signal intensity. As a result, the control unit 200 can perform MAP estimation processing that requires a large amount of calculation limitedly to a necessary region, and can generate a high-contrast image in a short time.

  As described above, the control unit 200 according to the present embodiment performs a plurality of tomographic images acquired by performing multiple optical coherence tomographic imaging using the measurement light controlled to scan substantially the same portion of the eye 140 to be examined. It is an image processing apparatus which processes the tomographic image produced | generated using. The control unit 200 includes an acquisition unit 220 that acquires information on a plurality of tomographic images. In addition, the control unit 200 generates a representative value at each pixel position of the tomographic image using information of the plurality of tomographic images, and generates a tomographic image using the representative value at each pixel position as the pixel value. Is included.

  The image generation unit 210 includes a region specifying unit 214 that uses a plurality of tomographic image information to specify a target region whose contrast is reduced in the tomographic image. In addition, the image generation unit 210 includes a calculation unit 215 that calculates the calculation value at the pixel position using the information of the plurality of tomographic images corresponding to the pixel position of the tomographic image among the information of the plurality of tomographic images. Further, the image generation unit 210 performs the MAP estimation process on the information of a plurality of tomographic images corresponding to the pixel positions among the information of the plurality of tomographic images at each pixel position of the target region, thereby obtaining an estimated value. An estimation unit 216 for estimation is provided. The tomographic image generation unit 213 of the image generation unit 210 generates a high-contrast tomographic image with the estimated value as a representative value in the target region and the calculated value as a representative value in a region other than the target region of the tomographic image.

  Regarding the specification of the target region, specifically, the region specifying unit 214 specifies the layer boundary of the fundus based on at least one tomographic image among the information of the plurality of tomographic images. Thereafter, the region specifying unit 214 specifies a target region from the specified layer boundary and the eye structure.

  Regarding the MAP estimation processing, specifically, the estimation unit 216 uses a predetermined probability density function for information on a plurality of tomographic images corresponding to each pixel position of the target region, and performs a plurality of corresponding to the pixel position. The probability density function of the information of the tomographic image is calculated. Thereafter, the estimation unit 216 calculates a product of probability density functions related to information of a plurality of tomographic signals corresponding to the same pixel position. Then, the estimation unit 216 determines information on a tomographic image that has the maximum probability in the product of probability density functions, and uses the determined value of information on the tomographic image as an estimated value.

  As described above, the control unit 200 according to the present embodiment performs the MAP estimation process only on the target area having a low signal intensity, and performs the averaging process on other areas such as the retina having a high signal intensity. This makes it possible to efficiently apply a MAP estimation process that requires a large amount of calculation, and to generate a tomographic image having a higher contrast than the averaged image in a shorter time than when the MAP estimation process is performed for all pixels. .

  In this embodiment, the luminance value of the tomographic image is used as the signal intensity in the addition averaging process, the area specifying process, and the MAP estimation process. However, the signal intensity used in these processes is not limited to the luminance value. Since the signal used for these processes may be a signal corresponding to each pixel position, it may be a tomographic signal based on a signal obtained by performing Fourier transform on the interference signal acquired by the OCT optical system. The tomographic signal includes, for example, a signal obtained by subjecting the interference signal to Fourier transform, a signal obtained by subjecting the signal after Fourier transform to arbitrary signal processing for image generation, and a signal such as a luminance value based thereon. In this specification, a set of tomographic signals corresponding to one tomographic image is referred to as one set of tomographic signals, and a set of tomographic signals corresponding to a plurality of tomographic images is referred to as a plurality of sets of tomographic signals.

  In the present embodiment, an example in which the fundus oculi Er is photographed is shown. However, the photographing target portion is not limited to the fundus oculi Er, and may be an anterior eye segment. Furthermore, in the present embodiment, a configuration has been described in which a region corresponding to a layer structure preset as a region with low signal intensity is specified as a target region to which MAP estimation is applied. However, the setting of the layer structure and organization corresponding to the region specified as the target region is not limited to this, and may be arbitrarily set according to a desired configuration. For example, it may be set by the examiner using a GUI, or may be set by selecting a recommended mode corresponding to a specific disease.

  In this embodiment, the region specifying unit 214 specifies the target region based on the average tomographic image, and the tomographic image generating unit 213 uses the estimated value based on the MAP estimation for the target region as a representative value to overwrite the pixel value. Contrast tomographic images were generated. However, the configuration for generating a high-contrast tomographic image is not limited to this. For example, a target region is specified based on at least one tomographic image among the acquired N tomographic images, and pixel values in the target region are estimated only by performing MAP estimation processing without performing averaging processing. The value may be a representative value.

(Example 2)
Hereinafter, the OCT apparatus according to the second embodiment will be described with reference to FIGS. Since the OCT apparatus according to the present embodiment has the same configuration as that of the OCT apparatus 1 according to the first embodiment, the same reference numerals are used for the constituent elements and the description thereof is omitted. Hereinafter, the OCT apparatus according to the present embodiment will be described focusing on differences from the OCT apparatus 1 according to the first embodiment.

  In the OCT apparatus 1 according to the first embodiment, the target region for performing the MAP estimation is specified using the segmentation function. On the other hand, in the OCT apparatus according to the present embodiment, the target region for performing the MAP estimation is specified using the signal intensity for each pixel.

  Further, in the OCT apparatus according to Example 1, a high-contrast tomographic image was generated when the high-contrast mode was previously selected. In contrast, in the OCT apparatus according to the present embodiment, after the tomographic image generation unit 213 generates the average tomographic image, the display control unit 250 displays the average tomographic image on the display unit 300, and then generates the high-contrast tomographic image. The operation of whether or not to do is accepted.

  FIG. 10 is a flowchart of tomographic image capturing processing according to the present embodiment. The shooting process according to the present embodiment is the same as the shooting process according to the first embodiment except for steps S1001 to S1004. Therefore, the description of the same process is omitted. When the imaging process according to the present embodiment is started, processes similar to the imaging process according to the first embodiment are performed in steps S501 and S502, and an interference signal is acquired in step S502.

  When the interference signal is acquired, in step S1001, the tomographic image generation unit 213 generates N tomographic images in the same manner as in step S503 of the imaging process according to the first embodiment. Here, in step S1001, the storage unit 240 further stores the generated N tomographic images for the post processing in the high contrast mode. Thereafter, processing similar to steps S504 and S505 of the imaging processing according to the first embodiment is performed, and in step S505, the tomographic image generation unit 213 generates an average tomographic image.

  When the average tomographic image is generated, the display control unit 250 displays the average tomographic image stored in the storage unit 240 on the display unit 300 in step S1002. FIG. 11 shows an example of a report screen 1100 displayed on the display unit 300 at this time. The report screen 1100 shows a post processing button 1101 in a high contrast mode along with the average tomographic image. When the examiner presses the post processing button 1101 in the high contrast mode using an input device (not shown), the post processing in the high contrast mode is turned on.

  Next, in step S1003, the tomographic image generation unit 213 determines whether or not the post processing button 1101 in the high contrast mode is pressed and the post processing in the high contrast mode is turned on. If the tomographic image generation unit 213 determines that the post processing in the high contrast mode is not turned on, the imaging processing is terminated.

  On the other hand, when the tomographic image generation unit 213 determines that the post processing in the high contrast mode is turned on, the processing proceeds to step S1004. In step S1004, the region specifying unit 214 specifies a target region using the histogram of the average tomographic image generated in step S505. In the present embodiment, a 10th percentile pixel value of all the pixels included in the average tomographic image is used as a threshold, and an area including pixels having pixel values equal to or lower than the threshold is specified as a target area. Thereby, the tomographic image generation unit 213 can specify a region including a pixel with low signal intensity as a target region.

  In this embodiment, the 10th percentile is used as the threshold value, but the threshold value setting is not limited to this. The threshold value may be arbitrarily set according to a desired configuration. As the threshold value, for example, a pixel value of an arbitrary percentile such as the 20th percentile may be set, or a value half the value obtained by averaging the pixel values of all the pixels included in the average tomographic image may be set.

  When the target area is specified in step S1004, the process proceeds to step S508. Since the processing after step S508 is the same as the imaging processing according to the first embodiment, description thereof is omitted.

  In the OCT apparatus according to the present embodiment, the region specifying unit 214 specifies a region at a pixel position having a calculated value lower than a threshold set using a calculated value calculated for each pixel position as a target region. Also in the OCT apparatus according to the present embodiment, the MAP estimation process is performed only on the region where the signal intensity is low, and the MAP estimation process that requires a large amount of calculation is efficiently applied, which is higher than the averaged image. A contrast tomographic image can be generated in a short time.

  Note that the region specifying unit 214 sets a threshold using one tomographic image among a plurality of tomographic images in the same manner as in the present embodiment, and selects a region having a tomographic signal value lower than the set threshold as the target region. May be specified.

  In the OCT apparatus according to the present embodiment, the high contrast mode can be selected by the examiner instructing from the GUI as necessary after displaying the average tomographic image. Therefore, the examiner can generate a high-contrast tomographic image after confirming that the average tomographic image has a low contrast. Accordingly, unnecessary calculation time can be avoided when the average tomographic image is sufficient depending on the usage of the image, and the operability and convenience of the OCT apparatus can be improved. The post processing in the high contrast mode can also be applied to the OCT apparatus 1 according to the first embodiment.

(Example 3)
Hereinafter, the OCT apparatus according to the third embodiment will be described with reference to FIG. Since the OCT apparatus according to the present embodiment has the same configuration as that of the OCT apparatus according to the second embodiment, the same reference numerals are used for the components and the description thereof is omitted. Hereinafter, the OCT apparatus according to the present embodiment will be described focusing on differences from the OCT apparatus according to the second embodiment.

  In the OCT apparatus according to the second embodiment, the storage unit 240 stores all of the N tomographic images generated in step S1001 for the post processing in the high contrast mode. Normally, when generating a tomographic image, data is handled in floating point 32 bits or 64 bits, and highly accurate calculation is performed. However, in the case where many tomographic images are always stored, if the data having such a high number of bits is stored, a large-capacity data storage device is required as the storage unit 240, which causes an increase in cost.

  For this reason, in this embodiment, the pixel value in each pixel of the tomographic image is subjected to gradation conversion, stored as data having a low number of bits, and processing for returning the gradation of the data is performed when performing MAP estimation.

  FIG. 12 is a flowchart of the photographing process according to the present embodiment. Since the shooting process according to the present embodiment is the same as the shooting process according to the second embodiment except for steps S1201 and S1202, the description of the same process is omitted. When the imaging process according to the present embodiment is started, the imaging process according to the second embodiment is performed in steps S501 to S1002. If the average tomographic image is displayed on the display unit 300 in step S1002, the process proceeds to step S1201.

  In step S <b> 1201, the tomographic image generation unit 213 performs gradation conversion on the N pieces of tomographic images stored in the storage unit 240. Specifically, the tomographic image generation unit 213 windows the pixel values of the tomographic image using the minimum and maximum pixel values in the tomographic image as the window width, and the pixel values included in the window width Is converted to 8-bit data by performing linear tone conversion. The storage unit 240 stores N tomographic images whose pixel values are tone-converted, instead of N tomographic images before tone-conversion. Regarding the tomographic image after gradation, since the above Equation 1 is used when performing the MAP estimation in step S508, it is necessary to return to the gradation used when generating the tomographic image. For this reason, the storage unit 240 also stores the conditions (such as the window width and the gradation ratio) used when the pixel values are converted to 8-bit data together with the tomographic image after gradation. In step S1201, when the storage unit 240 stores the tomographic image after gradation conversion, the process proceeds to step S1003.

  In step S1003, if the tomographic image generation unit 213 determines that the post processing in the high contrast mode is turned on, the process proceeds to step S1004. In step S1004, similarly to the imaging process according to the second embodiment, when the area specifying unit 214 specifies the target area based on the average tomographic image, the process proceeds to step S1202.

  In step S1202, the estimation unit 216 returns the gradations of the N tomographic images that have undergone gradation conversion in accordance with the stored gradation conversion conditions. Thereafter, in step S508, the estimation unit 216 performs MAP estimation processing according to the second embodiment using the N tomographic images whose gradations have been returned. Since the processing after step S508 is the same as the imaging processing according to the first embodiment, description thereof is omitted.

  As described above, when post-processing in the high contrast mode, it is necessary to always store a plurality of tomographic images used for MAP estimation. On the other hand, in the present embodiment, the storage unit 240 is configured to store a plurality of tomographic images that have been tone-converted into 8-bit information. By performing tone conversion on the tomographic image data to be stored, the image capacity to be stored can be suppressed and the cost can be suppressed.

  In this embodiment, tone conversion of N tomographic images is performed after the display of the average tomographic image, but the timing of tone conversion is not limited to this. For example, after generating N tomographic images, gradation conversion may be performed in parallel with the process of averaging the average tomographic images of N tomographic images, and the tomographic image after gradation conversion may be stored. In this case, after generating the average tomographic image, the storage unit 240 deletes the data of N pieces of tomographic images that have not been subjected to gradation conversion and used to generate the average tomographic image.

  In this embodiment, the tomographic image obtained by gradation-converting the pixel values related to all the pixels of the N tomographic images is stored, but the data after gradation conversion to be stored is not limited to this. Since the data after gradation conversion is stored for use in the MAP estimation process, only the data of the target area on which the MAP estimation process is performed may be stored. In this case, the target area setting process in step S1004 is performed prior to the gradation conversion process in step S1201. Thereafter, the pixel value of the target region is converted into 8-bit data by gradation conversion processing and stored in the storage unit 240.

  In addition, when the tone conversion is performed only on the pixel value of the target region having a low signal intensity, the window width of the windowing may be determined according to the target region specifying method. For example, the threshold value for specifying the target region may be the maximum value, and the maximum value and the minimum pixel value of the pixels in the tomographic image may be the window width. Alternatively, the pixel values of the tomographic image may be converted into a histogram, and the maximum value and minimum value of the lower 10% of the pixel values may be determined as the window width. In this case, the range of pixel values for determining the window width may be set to an arbitrary number such as the lower 20% or 30% according to the desired configuration. In this case, at the time of gradation conversion, it is possible to enrich the gradation in the area where the signal intensity is low, and the area of the data to be stored is limited, so that the image capacity to be stored can be further suppressed. Further, at the time of gradation conversion, gradation conversion may be performed not by linear conversion but by gamma correction so that gradation is rich in a region where the signal intensity is low.

(Modification of Examples 1 to 3)
In the first to third embodiments, the region specifying unit 214 specifies a target region on which the MAP estimation process is performed based on a threshold value obtained from the segmentation process or the histogram. On the other hand, the region specifying unit 214 may specify the target region based on the shape of the histogram of pixel values included in the tomographic image. FIG. 13 shows an example of a histogram of pixel values included in a tomographic image. In FIG. 13, the horizontal axis indicates the luminance value that is a pixel value, the vertical axis regarding the solid line graph indicates the number of pixels having the luminance value, and the vertical axis regarding the dotted line graph indicates the integrated value of the frequency.

  When the region specifying unit 214 determines the target region based on the shape of the histogram, for example, the luminance value of the percentile corresponding to the frequency integrated value for the luminance value at which the first peak appears when viewed from a low luminance value in the histogram. Can be set as a threshold value. In the example shown in FIG. 13, the frequency integrated value corresponding to the first peak indicated by the alternate long and short dash line is about 21%. Is specified as the target area. Note that any other threshold setting method based on the shape of the histogram for specifying a low signal strength region may be applied.

  Also in this modification, it is possible to specify a region having a low signal intensity as a target region. For this reason, the MAP estimation process is performed only on the region where the signal intensity is low. The MAP estimation process that requires a large amount of calculation is efficiently applied, and a tomographic image having a higher contrast than the averaged image can be shortened. Can be generated in time.

  In the first to third embodiments, since the signal strength after the averaging is higher than the true signal strength in the region where the signal strength is low, the MAP estimation process is applied to the region. However, the process in which the signal strength is higher than the true signal strength in the region where the signal strength is low is not limited to the averaging. For example, similar problems may occur in median value calculation mode, mode value calculation processing, maximum value calculation processing, and the like. Therefore, the present invention can also be applied to an image processing method for performing noise reduction by performing such processing on a plurality of tomographic images.

(Embodiment 2)
As described above, in the MAP estimation process described in Non-Patent Document 1, it is assumed that the signal intensity distribution of the OCT acquired with respect to the true signal intensity is based on the signal characteristics and the noise characteristics, and the distribution of a large number of signal acquisition results. The true signal strength at which the appearance probability of the signal based on is maximized is estimated. For this reason, for example, a result that is not correctly measured in the signal acquisition result due to an external factor such as the eye moving during measurement and the intensity of the reflected signal from the eye changing, or the amount of noise changing due to disturbance. If it is included, it leads to fluctuation of the estimation result.

  Here, when the measured signal strength is large with respect to the change in signal strength caused by an external factor, the influence is very small. However, when the measured signal strength is small, a minute change in signal strength due to an external factor may cause a large fluctuation in the estimation result. As a result, for example, when a tomographic image of the fundus is taken, roughness may be conspicuous on the image, particularly in the vitreous body, sclera, and choroid, which are regions with low signal intensity.

  Therefore, Embodiment 2 of the present invention provides an optical coherence tomographic imaging apparatus capable of generating an image having higher contrast and lower noise than that of conventionally used image averaging processing.

  Before describing an example of the present embodiment, the present embodiment will be schematically described. In the present embodiment, in view of the above-described problem that occurs for the first time in a high-contrast tomographic image obtained by performing the MAP estimation process, a process for making the roughness inconspicuous is performed. More specifically, a high-contrast image is generated by performing a MAP estimation process on the OCT tomographic signal, and then a process for making the roughness inconspicuous is performed on the high-contrast image. Provide fewer tomographic images.

(Example 4)
Hereinafter, an optical coherence tomographic imaging apparatus (OCT apparatus) according to Example 4 of the present embodiment will be described with reference to FIGS. Since the OCT apparatus according to the present embodiment has the same configuration as that of the OCT apparatus according to the first embodiment except for the control unit 1400, the same reference numerals are used for the constituent elements and the description thereof is omitted. Hereinafter, the OCT apparatus according to the present embodiment will be described focusing on differences from the OCT apparatus according to the first embodiment.

(Configuration of control unit 1400)
The configuration of the control unit 1400 will be described with reference to FIG. FIG. 14 is a block diagram illustrating a schematic configuration of the control unit 1400.

  The control unit 1400 is provided with an image generation unit 1410, an acquisition unit 220, a drive control unit 230, a storage unit 240, and a display control unit 250. Note that the acquisition unit 220, the drive control unit 230, the storage unit 240, and the display control unit 250 provided in the control unit 1400 are the same as those in the first embodiment, and thus the description thereof is omitted.

  The image generation unit 1410 includes an anterior eye image generation unit 211, a fundus image generation unit 212, a tomographic image generation unit 213, a calculation unit 215, an estimation unit 216, and a filter unit 1417. Note that the anterior eye image generation unit 211, the fundus image generation unit 212, the tomographic image generation unit 213, the calculation unit 215, and the estimation unit 216 provided in the image generation unit 1410 are the same as the respective components of the first embodiment. The description is omitted.

  The filter unit 1417 applies a noise removal filter to the high-contrast tomographic image based on the estimated value estimated by the estimation unit 216 to reduce noise in the high-contrast tomographic image. Here, the noise removal filter may be a filter used to reduce noise in an image. As the noise removal filter, for example, known filters such as an expansion / contraction filter, a moving average filter, a Gaussian filter, a weighting filter, a bilateral filter, and a median filter can be used. In this embodiment, a moving average filter is used. Note that the filter parameters, the weighting direction, and the like, which are filter parameters of the noise removal filter, can be arbitrarily set according to a desired configuration.

  Note that the filter unit 1417 can also be configured by a module executed by the CPU or MPU of the control unit 1400. The filter unit 1417 may be configured by a circuit or the like that realizes a specific function such as an ASIC.

(Tomographic imaging process)
Hereinafter, with reference to FIGS. 15 to 18, a tomographic image capturing process according to the present embodiment will be described. FIG. 15 shows an example of a control / image display GUI preview screen 1500 displayed on the display unit 300 according to the present embodiment. The control / image display GUI report screen displayed on the display unit 300 is the same as the report screen 406 according to the first embodiment, and a description thereof will be omitted.

  A preview screen 1500 shown in FIG. 15 is a screen that is displayed to perform an instruction to start imaging, alignment adjustment of the imaging optical system 100, adjustment of the position of a part to be imaged, and the like. In this embodiment, a function of generating a high-contrast tomographic image by pressing a high-contrast mode button 405 on the preview screen 1500 shown in FIG. 15 and turning on the high-contrast mode using an input device (not shown). Can be used. Similarly, a function for generating a high-contrast and low-noise tomographic image can be used by pressing the roughness suppression mode button 1501 and turning on the roughness suppression mode.

  In this embodiment, after preparing for photographing on the preview screen 1500 shown in FIG. 15, the fundus Er of the subject is photographed using the OCT apparatus 1. Note that the procedures for starting shooting preparation and shooting are the same as those according to the first embodiment, and a description thereof will be omitted.

  FIG. 16 shows a flowchart of tomographic image capturing processing according to this embodiment. In addition, since the noise signal acquisition to the alignment process and the ON / OFF determination of the high contrast mode according to the present embodiment are the same as the same process according to the first embodiment, the description is omitted using the same reference numerals.

  In step S504, when the alignment of tomographic images is completed, the process proceeds to step S506. In step S506, the tomographic image generation unit 213 determines whether or not the high contrast mode is turned on. If the tomographic image generation unit 213 determines in step S506 that the high contrast mode has not been turned ON, the process proceeds to step S1605.

  In step S1605, the calculation unit 215 calculates a calculation value of N pixel values for each pixel with respect to the aligned N tomographic images, and calculates the calculation value calculated by the tomographic image generation unit 213 for each pixel. An averaged tomographic image (average tomographic image) is generated using the representative value. The storage unit 240 stores the generated average tomographic image. In the present embodiment, the calculation unit 215 calculates the calculated value by averaging the luminance values at the pixel positions in the N tomographic images corresponding to the pixel positions of the generated average tomographic image. Then, the tomographic image generation unit 213 generates an average tomographic image using the calculated calculation value as the luminance value at each pixel position. Note that, as described above, the calculated value is not limited to an arithmetic average value (arithmetic average value), and may be a median value, a mode value, a maximum value, or the like.

  Thereafter, in step S1604, the display control unit 250 displays the average tomographic image stored in the storage unit 240 in step S1605 on the display unit 300, and ends the imaging process.

  On the other hand, if the tomographic image generation unit 213 determines in step S506 that the high contrast mode is ON, the process proceeds to step S1601.

  In step S1601, the estimation unit 216 performs MAP estimation processing on the N tomographic images that have been aligned, estimates an estimated value at each pixel position, and uses the estimated value estimated by the tomographic image generation unit 213. A contrast tomographic image is generated. Hereinafter, the MAP estimation processing according to the present embodiment will be described with reference to FIG. FIG. 17 shows a flowchart of the MAP estimation process according to this embodiment. In the MAP estimation processing according to the present embodiment, the processing related to the calculation of the noise variance, the probability density map, the probability density function, the product of the probability density function, and the extraction of the true signal strength is performed by the MAP estimation according to the first embodiment. Since it is the same as the process, the same reference numerals are used and description thereof is omitted.

  When the probability density map is calculated in step S702, the process proceeds to step S1701. In step S1701, the estimation unit 216 sets an initial position of the pixel position (coordinates (x, z)) when starting the MAP estimation process. Here, coordinates (x, z) indicate coordinates in a high-contrast tomographic image (high-contrast tomographic image) to be generated, x indicates a position in the B-scan direction in the tomographic image, and z indicates a depth in the tomographic image. Indicates the position of the direction. Each coordinate in the high-contrast tomographic image to be generated corresponds to each coordinate in the average tomographic image and the N tomographic images that have been aligned. In this embodiment, the initial value of the pixel position is set to the coordinates (1, 1). Note that the initial value is not limited to this, and may be coordinates of an arbitrary pixel position in the tomographic image.

  Next, in steps S704 to S706, the estimation unit 216 determines a parameter (true signal strength ν) having the maximum probability in the posterior probability function, as in the MAP estimation process according to the first embodiment. When the true signal strength ν is determined, the process proceeds to step S1702.

  In step S1702, the tomographic image generation unit 213 uses the estimated value (determined true signal intensity ν) estimated by the MAP estimation processing by the estimation unit 216 at the pixel position corresponding to the coordinates (x, z) as a representative value. And adopted as the pixel value.

  Thereafter, in step S1703, the estimation unit 216 determines whether or not estimated values have been estimated for all pixel positions. If the estimation unit 216 determines that the estimated values are not estimated for all pixel positions, the process proceeds to step S1704. In step S1704, the estimation unit 216 moves the pixel position (coordinates (x, z)) from which the estimated value is estimated, and then returns the process to step S704.

  If it is determined in step S1703 that the tomographic image generation unit 213 has determined pixel values based on the estimated values for all pixel positions, the process proceeds to step S1602. The tomographic image generation unit 213 can generate a high-contrast tomographic image having pixel values based on the estimated values obtained by the MAP estimation process by determining pixel values based on the estimated values for all pixel positions. The storage unit 240 stores the generated high contrast tomographic image.

  In step S1602, the tomographic image generation unit 213 determines whether or not the roughness suppression mode is ON. If the tomographic image generation unit 213 determines in step S1602 that the roughness suppression mode is not turned on, the process proceeds to step S1604. In step S1604, the display control unit 250 displays the high-contrast tomographic image stored in the storage unit 240 on the display unit 300, and the imaging process ends.

  On the other hand, if the tomographic image generation unit 213 determines in step S1602 that the roughness suppression mode is ON, the process proceeds to step S1603.

  In step S1603, the filter unit 1417 applies a moving average filter that is a noise removal filter to the high-contrast tomographic image stored in the storage unit 240, thereby reducing noise in the high-contrast tomographic image.

  FIG. 18A shows an example of a high-contrast tomographic image generated by the MAP estimation process described above. FIG. 18C shows a one-line profile of a high-contrast tomographic image corresponding to the alternate long and short dash line in FIG. In the high-contrast tomographic image shown in FIG. 18A, it can be seen that noise 1801 having a strong contrast is generated in the vitreous region. In particular, when a one-line profile of a high-contrast tomographic image corresponding to an A-scan image is seen, it can be seen that noise appears with high contrast as shown in FIG.

  In order to suppress this noise, the filter unit 1417 performs a filtering process on the high-contrast tomographic image using a 3 × 3 pixel moving average filter. Specifically, the filter unit 1417 performs filter processing by convolving a high-contrast tomographic image and a 3 × 3 moving average filter. By this processing, it is possible to generate a tomographic image in which noise of the high contrast tomographic image is reduced.

  FIG. 18B shows an example of the generated high-contrast and low-noise tomographic image. FIG. 18D shows a one-line profile of a high-contrast tomographic image corresponding to the alternate long and short dash line in FIG. In FIG. 18B, noise 1801 generated in the high contrast tomographic image shown in FIG. 18A is reduced. In particular, looking at the profile of one line of the high-contrast tomographic image corresponding to the A scan image, it can be seen that noise is reduced as shown in FIG.

  When the filter unit 1417 generates a high-contrast and low-noise tomographic image, the storage unit 240 stores the generated high-contrast and low-noise tomographic image, and the process proceeds to step S1604.

  In step S1604, the display control unit 250 displays the high-contrast and low-noise tomographic image stored in the storage unit 240 on the display unit 300, and the imaging process ends.

  As described above, the control unit 1400 applies the noise removal filter to the high-contrast tomographic image generated by performing the MAP estimation process. Thereby, the control unit 1400 can generate a high-contrast and low-noise tomographic image.

  As described above, the image generation unit 1410 of the control unit 1400 according to this embodiment includes the filter unit 1417 that applies the noise removal filter to the high-contrast tomographic image generated using the estimated value as the representative value. Prepare. The control unit 1400 according to the present embodiment applies a noise removal filter to the high-contrast tomographic image generated based on the estimated value obtained by the MAP estimation process. Thereby, a tomographic image with higher contrast and lower noise than the average tomographic image can be generated.

  In this embodiment, the luminance value of the tomographic image is used as the signal intensity in the addition averaging process and the MAP estimation process. However, the signal intensity used in these processes is not limited to the luminance value. Since the signal used for these processes may be a signal corresponding to each pixel position, it may be a tomographic signal based on a signal obtained by performing Fourier transform on the interference signal acquired by the OCT optical system. The tomographic signal includes, for example, a signal obtained by subjecting the interference signal to Fourier transform, a signal obtained by subjecting the signal after Fourier transform to arbitrary signal processing for image generation, and a signal such as a luminance value based thereon.

  In the present embodiment, an example in which the fundus oculi Er is photographed is shown. However, the photographing target portion is not limited to the fundus oculi Er, and may be an anterior eye segment.

(Example 5)
In the fourth embodiment, the filtering process is performed on all the pixels of the high-contrast tomographic image. However, in this embodiment, the filtering process is performed only on a specific region. In the filter processing of the fourth embodiment, the noise removal filter is also applied to the retina. In general, a noise removal filter such as a moving average filter tends to reduce spatial resolution. Therefore, there is a case where a high-definition tomographic image for use in observing a subject in detail cannot be generated.

  On the other hand, there are three types of regions, vitreous body, choroid, and sclera, that are roughly divided into regions where the signal intensity that causes roughness during fundus observation is low. Of these areas, especially in the vitreous and sclera areas, it is known that there are no anatomically minute structures, so detailed observations can be made even if the spatial resolution is reduced. Is possible. Therefore, detailed observation can be performed for these regions even if a noise removal filter is applied.

  Therefore, in this embodiment, when a high-contrast tomographic image is generated, the filter process is applied only to those regions where the signal intensity that is likely to generate noise is low. More specifically, the fundus tomographic image in the tomographic image is separated into anatomical layers, and the noise removal filter is applied only to a region known as a region with low signal intensity. Thereby, it is possible to obtain a high-contrast tomographic image in which roughness (noise) in a region where the signal intensity is low is reduced.

  Hereinafter, the OCT apparatus according to the fifth embodiment will be described with reference to FIGS. In the OCT apparatus according to the present embodiment, noise removal filter processing is applied only to the vitreous body, which is a region having a low signal intensity obtained when a fundus tomographic signal is acquired, to reduce noise in a high-contrast tomographic image. Constituent elements of the OCT apparatus according to the present embodiment are the same as those of the OCT apparatus according to the fourth embodiment, and description thereof is omitted using the same reference numerals. Hereinafter, the OCT apparatus according to the present embodiment will be described focusing on differences from the OCT apparatus according to the fourth embodiment.

  FIG. 19 illustrates a schematic configuration of a control unit 1900 (image processing apparatus) according to the present embodiment. The image generation unit 1910 of the control unit 1900 includes an area specifying unit 1918 in addition to the anterior eye image generation unit 211 to the filter unit 1417.

  Based on the tomographic image generated by the tomographic image generation unit 213, the region specifying unit 1918 specifies a target region having a low signal intensity to which the noise removal filter should be applied. Note that the region specifying unit 1918 may specify the target region based on the Fourier-transformed signal acquired by the acquiring unit 220 or the like.

  Next, photographing processing according to the present embodiment will be described with reference to FIG. FIG. 20 is a flowchart of tomographic image capturing processing according to the present embodiment. The shooting process according to the present embodiment is the same as the shooting process according to the fourth embodiment except that the target area to which the filter is applied is specified and the filter is applied only to the specified target area. Is omitted.

  If the tomographic image generation unit 213 determines in step S1602 that the roughness suppression mode is ON, the process proceeds to step S2001.

  In step S2001, the region specifying unit 1918 performs segmentation processing on the high-contrast tomographic image generated based on the estimated value, and specifies a vitreous region, which is a region having a low signal intensity, as a target region. The specific process is the same as the process performed by the region specifying unit 214 according to the first embodiment, and a description thereof will be omitted.

  In the present embodiment, the region specifying unit 1918 extracts an internal limiting membrane (ILM) by segmentation processing. The area specifying unit 1918 determines that the side opposite to the retina (pupil side) of the ILM boundary is a vitreous body, and the area on the pupil side from the ILM boundary (area 2101 shown in FIG. 21) is known as a low signal intensity area. Identified as the vitreous region. Thereby, the area specifying unit 1918 specifies the vitreous area as a target area to which the noise removal filter is applied.

  The region specifying unit 1918 may perform segmentation processing on at least one of the N tomographic images that have been aligned, specify a region having a low signal intensity, and set the region as a target region. The segmentation process may be performed using any known method other than the above method. Furthermore, in this embodiment, the vitreous region is set as the target region. However, for example, an arbitrary region such as a sclera region known as a low signal intensity region may be specified as the target region. In this case, the region specifying unit 1918 can specify a layer such as a sclera by the segmentation process, and can set the layer as a target region. Further, the target region is not limited to one of these regions, and may be, for example, both the vitreous region and the sclera region.

  When the target region is specified, in step S2002, the filter unit 1417 applies a noise removal filter to the target region. More specifically, the filter unit 1417 performs convolution using a Gaussian filter with 5 × 5 pixels and σ = 0.5 only in the target region of the high-contrast tomographic image generated based on the estimated value, and the filter Perform the process. As a result, it is possible to generate a high-contrast and low-noise tomographic image in which the noise removal filter is applied only to a target region with low signal intensity. The storage unit 240 stores the generated tomographic image. Thereafter, in step S1604, the display control unit 250 displays the generated high-contrast and low-noise tomographic image on the display unit 300. The noise removal filter to be applied is not limited to the Gaussian filter, and may be any of the above-described filters such as a moving average filter and a weighting filter. Also, the filter parameters of the noise removal filter such as the filter size may be arbitrarily set.

  As described above, the image generation unit 1910 according to the present embodiment includes the region specifying unit 1918 that specifies a target region in a tomographic image using information of a plurality of tomographic images. More specifically, the region specifying unit 1918 specifies an anatomical layer (layer boundary) of the fundus based on the high-contrast tomographic image generated using the estimated value as a representative value. Thereafter, the area specifying unit 1918 specifies the target area from the specified layer and eye structure. The filter unit 1417 applies the noise removal filter only to the target region specified by the region specifying unit 1918.

  As a result, the OCT apparatus according to the present embodiment can obtain a fundus tomographic image with high contrast and suppressed noise in the target region without reducing the spatial resolving power of the retina structure. In addition, the vitreous body and sclera regions having low signal intensity are regions that do not have a minute structure, so that even if a noise removal filter is applied, detailed observation can be performed.

  In this embodiment, the region specifying unit 1918 specifies the anatomical layer of the fundus based on the high-contrast tomographic image generated using the estimated value as a representative value. However, the region specifying unit 1918 may specify the anatomical layer of the fundus based on the information of at least one tomographic image among the information of the plurality of tomographic images.

  In the present embodiment, a configuration has been described in which a region corresponding to a layer structure preset as a low signal intensity region is specified as a target region to which the noise removal filter is applied. However, the setting of the layer structure and organization corresponding to the region specified as the target region is not limited to this, and may be arbitrarily set according to a desired configuration. For example, it may be set by the examiner using a GUI, or may be set by selecting a recommended mode corresponding to a specific disease.

(Modification of Example 5)
In a region where the signal intensity such as the retina is sufficiently high during fundus observation, there is no difference between the results of the averaging process and the MAP estimation process. For this reason, the retinal area may be averaged without performing the MAP estimation process, and the MAP estimation process and the noise removal filter may be performed only on the target area having a low signal strength. Thereby, it is possible to limit the calculation area by limiting the area where the MAP estimation process is performed, shorten the calculation time, and reduce the roughness in the high-contrast tomographic image.

  Here, a modification of the fifth embodiment will be described with reference to FIG. In the fifth embodiment, only the region where the signal intensity is low is specified and the filter process is performed, but the MAP estimation process is performed on the entire image. In this modification, this MAP estimation process is also performed only on the same area as the area where the filter process is performed.

  FIG. 22 is a flowchart of tomographic image capturing processing according to a modification of the fifth embodiment. Hereinafter, the same processing as the tomographic image capturing processing according to the fifth embodiment will be omitted, and the description will focus on the differences.

  In this modification, after the alignment process is performed in step S504, the process proceeds to step S2201. In step S2201, similarly to step S1605 of the process according to the fifth embodiment, the calculation unit 215 calculates a calculation value, and the tomographic image generation unit 213 generates an average tomographic image using the calculated value. The storage unit 240 stores the generated average tomographic image.

  Thereafter, the process proceeds to step S506. If the tomographic image generation unit 213 determines in step S506 that the high contrast mode has not been turned ON, the process proceeds to step S1604, and the display control unit 250 causes the display unit 300 to display the average tomographic image. . On the other hand, if the tomographic image generation unit 213 determines that the high contrast mode is ON, the process proceeds to step S2202.

  In step S2202, the area specifying unit 1918 performs the area specifying process in the same manner as in step S2001 of the fifth embodiment. Here, in the present embodiment, the region specifying unit 1918 performs region specifying processing on the average tomographic image generated in step S2201, and specifies a target region to which the MAP estimation processing and filter processing are applied.

  When the target area is specified, in step S2203, the estimation unit 216 performs MAP estimation processing for all the pixels in the target area. Here, the MAP estimation process in the present embodiment is the same as the MAP estimation process shown in FIG. 17, but since the target to be processed is a pixel in the target region, the estimation unit 216 determines the initial position in step S1701. The pixel position in the target area is determined. The initial position may be an arbitrary position as long as it is a position within the target region. For example, the estimation unit 216 sets the coordinates having the smallest values of x and z in the target region as the initial position. In step S1702, the tomographic image generation unit 213 overwrites the pixel value with the signal intensity extracted in step S708 as a representative value at the pixel position of the average tomographic image corresponding to the coordinates (x, z). In step S1703, the estimation unit 216 determines whether or not processing has been performed for all pixel positions in the target region. If there is an unprocessed pixel position, the pixel position is unprocessed in step S1704. To the pixel position.

  As described above, in the present modification, the tomographic image generation unit 213 generates a tomographic image using the estimated value as a representative value in a target region of the tomographic image and the calculated value as a representative value in a region other than the target region. Accordingly, since the MAP estimation process is performed only on the target region, the number of pixels on which the MAP estimation process is performed is reduced, and the calculation time can be shortened. As described above, the calculated value can be any one of the arithmetic average value, the median value, the mode value, and the maximum value.

  In this modification, the region specifying unit 1918 specifies the target region based on the average tomographic image, and the tomographic image generating unit 213 uses the estimated value based on the MAP estimation for the target region as a representative value to overwrite the pixel value. Contrast tomographic images were generated. However, the configuration for generating a high-contrast tomographic image is not limited to this. For example, the target region may be specified based on at least one of the N tomographic images that have been aligned. Also, pixel values in the target area may be subjected to only MAP estimation processing without performing averaging processing, and the estimated value may be used as a representative value. Note that the area specifying unit 1918 may specify a target area for performing MAP estimation using the signal intensity for each pixel, as in the second embodiment.

(Example 6)
In the OCT apparatus according to the fifth embodiment, the region specifying unit 1918 specifies a region on which filter processing is performed using layer boundary extraction processing. On the other hand, in the OCT apparatus according to the sixth embodiment, a region having a low signal strength after the MAP estimation process is extracted as a target region, and the extracted target region is subjected to filter processing.

  Hereinafter, the OCT apparatus according to this embodiment will be described with reference to FIGS. Constituent elements of the OCT apparatus according to the present embodiment for explaining the region specifying process according to the present embodiment are the same as those of the OCT apparatus according to the fifth embodiment, and thus the description thereof is omitted using the same reference numerals. Hereinafter, the OCT apparatus according to the present embodiment will be described focusing on differences from the OCT apparatus according to the fifth embodiment.

  FIG. 23A shows an example of a high-contrast tomographic image generated by the MAP estimation process. FIG. 23B shows a histogram of the high contrast tomographic image shown in FIG. FIGS. 23C and 23D show examples of target areas specified by the area specifying unit 1918. In FIG. 23B, the horizontal axis indicates the luminance value, the vertical axis of the histogram indicates the number (frequency) of pixels having the luminance value, and the vertical axis regarding the solid line graph indicates the integrated value of the frequency.

  In step S2001, the area specifying unit 1918 according to the present embodiment creates a histogram (FIG. 23B) for the high-contrast tomographic image generated based on the estimated value as shown in FIG. Thereafter, as shown in FIG. 23C, a region having a signal intensity (luminance value) lower than the entire 20th percentile in the histogram is specified as a target region 2301. More specifically, the pixel value (luminance value) of the pixels corresponding to the entire 20th percentile is set as a threshold value, and a pixel region having a luminance value lower than the threshold value is specified as the target region 2301. Note that the threshold is not limited to the 20th percentile, and can be set to an arbitrary value such as the 10th percentile or the 30th percentile according to the desired configuration.

  At this time, as shown in FIG. 23C, there may be a case where a minute region 2302 in which noise to be removed such as sesame salt noise is generated is not included in the target region. Therefore, the region specifying unit 1918 applies a known technique used when generating a mask with a binary image, such as binary image enlargement / contraction processing or region filling processing, to the specified target region 2301. be able to. Accordingly, as illustrated in FIG. 23D, the region specifying unit 1918 can set a region with a low signal strength to which the noise removal filter should be applied as the target region 2303.

  Thereafter, in step S2002, the filter unit 1417 performs high-contrast tomographic image filter processing on the set target region using a 3 × 3 moving average filter. Note that the type and filter parameters of the noise removal filter may be arbitrary as in the fifth embodiment.

  As described above, in the OCT apparatus according to the present embodiment, the region specifying unit 1918 specifies the region of the pixel position having a pixel value lower than the threshold set using the estimated value as the target region. As a result, similarly to the fifth embodiment, it is possible to obtain a fundus tomographic image in which the vitreous noise is suppressed with high contrast without lowering the spatial resolution of the structure of the retina.

  In this embodiment, the region specifying unit 1918 creates a histogram for the high-contrast tomographic image generated based on the estimated value, and sets a threshold value. However, the method of setting the threshold is not limited to this. For example, a histogram may be created for at least one tomographic image among N tomographic images that have been aligned, and a threshold value may be set from the histogram. Further, the area specifying process according to the present embodiment may be applied to the area specifying process in the modification of the fifth embodiment.

(Embodiment 3)
As described above, since high contrast of tomographic images by MAP estimation processing described in Non-Patent Document 1 is a statistical estimation technique, the more pixels (the number of tomographic images) used, the more reliable A high true value can be estimated.

  However, the subject must naturally keep his eyes open while taking a tomographic image of the fundus. Therefore, when taking a large number of tomographic images, the subject is extremely burdened. In addition, since it takes a long time to take a large number of tomographic images, the subject is burdened with time. In addition to this, if the imaging time becomes longer, there is a higher possibility that the imaging will fail, and there will be a further burden on the subject, such as the need to perform reimaging.

  Therefore, Embodiment 3 of the present invention provides an optical coherence tomographic imaging apparatus that generates a high-contrast tomographic image using less tomographic information than in the past.

  Before describing an example of the present embodiment, the present embodiment will be schematically described. In Non-Patent Document 1, the signal intensity of pixels at the same location in a large number of tomographic images is used as a sample, and a true value (signal intensity) is estimated from their probability density distribution. On the other hand, in this embodiment, in addition to the pixels at the same location in a plurality of tomographic images, the signal intensity of pixels around the pixel position is also treated as a sample, and the true value is estimated by increasing the number of samples. Hereinafter, such MAP estimation processing is referred to as range MAP estimation processing. In the range MAP estimation process according to the present embodiment, the number of samples increases by performing the MAP estimation process including surrounding pixels, so the number of tomographic images (number of tomographic information) required to obtain a high-contrast tomographic image. ) Can be reduced.

(Example 7)
Hereinafter, an optical coherence tomography apparatus (OCT apparatus) according to Example 7 of this embodiment will be described with reference to FIGS. Since the OCT apparatus according to the present embodiment has the same configuration as the OCT apparatus according to the first embodiment except for the control unit 2400, the same reference numerals are used for the constituent elements and the description thereof is omitted. Hereinafter, the OCT apparatus according to the present embodiment will be described focusing on differences from the OCT apparatus according to the first embodiment.

(Configuration of control unit 2400)
With reference to FIG. 24, the configuration of the control unit 2400 (image processing apparatus) will be described. FIG. 24 is a block diagram illustrating a schematic configuration of the control unit 2400. The control unit 2400 is provided with an image generation unit 2410, an acquisition unit 220, a drive control unit 230, a storage unit 240, and a display control unit 250. Note that the acquisition unit 220, the drive control unit 230, the storage unit 240, and the display control unit 250 provided in the control unit 2400 are the same as those in the first embodiment, and thus the description thereof is omitted.

  The image generation unit 2410 includes an anterior eye image generation unit 211, a fundus image generation unit 212, a tomographic image generation unit 213, a calculation unit 215, and an estimation unit 2416. Note that the anterior eye image generation unit 211, the fundus image generation unit 212, the tomographic image generation unit 213, and the calculation unit 215 provided in the image generation unit 2410 are the same as the components of the first embodiment, and thus the description thereof is omitted. To do.

  The estimation unit 2416 performs a range MAP estimation process to be described later on the plurality of tomographic images generated by the tomographic image generation unit 213, and estimates an estimated value of the true value of the interference signal. Note that the estimation unit 2416 may estimate the estimated value at each pixel position based on the signal after Fourier transform acquired by the acquisition unit 220 or the like.

  Note that the estimation unit 2416 can also be configured by a module executed by the CPU or MPU of the control unit 2400. In addition, the estimation unit 2416 may be configured by a circuit that implements a specific function such as an ASIC.

(Tomographic imaging process)
Hereinafter, with reference to FIGS. 25 to 27, a tomographic image capturing process according to the present embodiment will be described. Since the preview screen of the control / image display GUI and the report screen of the control / image display GUI according to the present embodiment are the same as the preview screen 400 and the report screen 406 according to the first embodiment, description thereof will be omitted. In addition, the procedure for starting shooting preparation and shooting is the same as those according to the first embodiment, and a description thereof will be omitted.

  FIG. 25 is a flowchart of tomographic image capturing processing according to the present embodiment. In addition, since the noise signal acquisition to the alignment process and the ON / OFF determination of the high contrast mode according to the present embodiment are the same as the same process according to the first embodiment, the description is omitted using the same reference numerals. In addition, the averaging process and the display process according to the present embodiment are the same as the same process according to the fourth embodiment, and thus the description thereof is omitted using the same reference numerals.

  In this embodiment, the number of times of imaging for acquiring the interference signal in step S502 is N = 10. However, the number of times of imaging is not limited to this. N, which is the number of times of imaging, may be an arbitrary number equal to or greater than 2 so that a desired number of tomographic images for performing the tomographic image range MAP estimation process and the addition averaging process can be obtained. However, in the range MAP estimation process, since the number of tomographic information (tomographic images) used for estimation value estimation can be reduced as compared with the conventional MAP estimation process, N is larger than the number of imaging required in the conventional MAP estimation process. It can be a small number.

  If the tomographic image generation unit 213 determines in step S506 that the high contrast mode has not been turned on, that is, has been turned off, the process proceeds to step S1605. In step S1605, the calculation unit 215 generates an averaged tomographic image (average tomographic image) as in the fourth embodiment, and in step S1604, the display control unit 250 causes the display unit 300 to display the average tomographic image. To complete the imaging process.

  On the other hand, if the tomographic image generation unit 213 determines in step S506 that the high contrast mode is ON, the process proceeds to step S2501.

  In step S2501, the estimation unit 2416 performs a range MAP estimation process on the aligned N tomographic images, estimates an estimated value at each pixel position, and uses the estimated value estimated by the tomographic image generation unit 213. A high-contrast tomographic image is generated. Here, the MAP estimation process itself in the range MAP estimation process is the same as that described in Non-Patent Document 1, but differs in that the signal intensity of surrounding pixels in addition to the target pixel is also used in the estimation process. .

  Hereinafter, the range MAP estimation process according to the present embodiment will be described with reference to FIG. FIG. 26 shows a flowchart of the range MAP estimation process according to this embodiment. In the range MAP estimation process according to the present embodiment, the process related to the calculation of the noise variance and the probability density map is the same as the MAP estimation process according to the first embodiment, and thus the description is omitted using the same reference numerals. .

  When the range MAP estimation process is started and the noise variance and probability density map are calculated in steps S701 and S702, the process proceeds to step S2601. In step S2601, the estimation unit 2416 sets an initial position of the pixel position (coordinates (x, z)) when starting the range MAP estimation process. In this embodiment, the initial value of the pixel position is set to the coordinates (1, 1). Note that the initial value is not limited to this, and may be coordinates of an arbitrary pixel position in the tomographic image.

  Further, in the following, pixels around the target pixel (x, z) (peripheral pixels) used in the range MAP estimation process according to the present embodiment are pixels above, below, left, and right of the target pixel, that is, coordinates (x−1, z). ), (X, z−1), (x, z + 1), and (x + 1, z) pixels. At this time, there are cases where surrounding pixels protrude outside the image at the peripheral edge of the tomographic image. In this case, it can be dealt with by reducing the number of pixels used for the range MAP estimation processing by the amount of the protruding pixels. Further, various boundary processing may be appropriately performed such that the peripheral pixels of the tomographic image are not used for final image display, and the range MAP estimation processing is not performed on these pixels.

  Note that the surrounding pixels of the target pixel used for the range MAP estimation process are not limited to the upper, lower, left, and right pixels of the target pixel. The surrounding pixels of the target pixel used for the range MAP estimation process may be pixels around the target pixel. The surrounding pixels may be, for example, 8 pixels around the target pixel in the 3 × 3 pixel centered on the target pixel, or 24 around the target pixel in the 5 × 5 pixel centered on the target pixel. It may be a pixel. The surrounding pixels may be pixels that are diagonally adjacent to the target pixel. In this specification, the surrounding pixels of the target pixel used in the range MAP estimation process are simply referred to as surrounding pixels, and the pixel positions of the surrounding pixels are referred to as pixel positions around the target pixel.

Next, in step S2602~ step S2604, the estimation unit 2416, a parameter a true signal strength ν for each signal strength _{a n} in the coordinates (x, z) of the target pixel 1~N th tomographic image The obtained probability density is extracted from the probability density map. Similarly, the estimation unit 2416 coordinates (x-1, z), (x, z-1), (x, z + 1), (x + 1, z) of the surrounding pixels of the first to Nth tomographic images. the probability density extracted from the probability density map for each signal strength a _n in. Thereafter, the estimation unit 2416 calculates a probability density function using the true signal intensity ν as a parameter from each of the extracted probability densities. Note that nmax in the figure corresponds to N.

In this embodiment, the estimation unit 2416 has coordinates (x, z), (x−1, z), (x, z−1), (x, z + 1), (x, z) of the nth tomographic image to be processed. x + 1, z) extracts the probability density for the signal strength a _n along the direction of the broken arrow of the probability density map shown in Figure 8 in. If there is a tomographic image for which processing has been determined to be canceled at the time of registration, the estimation unit 2416 extracts probability density for the number of tomographic images excluding the number of tomographic images for which processing has been determined to be stopped. To do. When estimating unit 2416 extracts the probability density of the true signal intensity [nu to the signal strength a _n in these coordinates N sheets, each calculated probability density function as a parameter the true signal strength [nu, the process to step S2605 Transition.

In step S2605, the estimation unit 2416 calculates the probability density function of the true signal strength ν with respect to the measured signal strengths of N × 5 (for the target pixel and surrounding pixels) from the signal strengths a ₁ to a _N at these coordinates. And the calculated product is used as a likelihood function. Since the product of the probability density function is the same as the product of the probability density function in the first embodiment, the description thereof is omitted. However, in the range MAP estimation process, the probability density function of the true signal intensity ν relating to the surrounding pixels is similarly multiplied by the probability density function of the true signal intensity ν relating to the target pixel, so that the probability density function for the target pixel is obtained. Is calculated.

  In step S2606, the true signal strength ν that maximizes the probability is determined in the posterior probability function (representative probability density function) that is the product of the likelihood function calculated by the estimation unit 2416 and the prior probability. Here, considering a uniform prior probability that a signal is always acquired as the prior probability, and assuming the prior probability as 1, the likelihood function becomes the posterior probability function as it is. The estimation unit 2416 determines a parameter (true signal strength ν) having the maximum probability in this posterior probability function.

More simply, in this embodiment, the estimation unit 2416 performs the probability density for each true signal strength ν with respect to the measured signal strengths a _{1 to} a _{N at} the target pixel and the surrounding pixels in steps S2602 to S2604. To extract. Thereafter, in step S2605, a product of probability density functions is calculated by multiplying each signal intensity by the probability density for each extracted true signal intensity ν. Then, the true signal strength ν that maximizes the probability is determined in the product of the calculated probability density functions. In this way, the determined true signal strength ν becomes an estimated value estimated by the estimating unit 2416.

  In step S2607, the tomographic image generation unit 213 represents the estimated value (determined true signal intensity ν) estimated by the range MAP estimation processing by the estimation unit 2416 at the pixel position corresponding to the coordinates (x, z). Determined as a value and adopted as a pixel value.

  Thereafter, in step S2608, the estimation unit 2416 determines whether estimated values have been estimated for all pixel positions. If the estimation unit 2416 determines that the estimated values are not estimated for all pixel positions, the process proceeds to step S2609. In step S2609, the estimation unit 2416 moves the pixel position (coordinates (x, z)) from which the estimated value is estimated, and then returns the process to step S2602.

  If it is determined in step S2608 that the tomographic image generation unit 213 has determined pixel values based on the estimated values for all pixel positions, the process proceeds to step S1604. The tomographic image generation unit 213 can generate a high-contrast tomographic image having pixel values based on the estimated values obtained by the range MAP estimation process by determining pixel values based on the estimated values for all pixel positions. The storage unit 240 stores the generated high contrast tomographic image.

  In step S1604, the display control unit 250 displays the high-contrast tomographic image stored in the storage unit 240 on the display unit 300, and the imaging process ends.

  As described above, the control unit 2400 can generate the high-contrast tomographic image while reducing the number of tomographic images to be captured by performing the range MAP estimation process.

  Further, by performing the range MAP estimation process, it is possible to reduce noise in the high-contrast tomographic image. Here, the noise reduction will be described with reference to FIGS.

  In normal MAP estimation processing, the distribution of OCT signal strength acquired with respect to the true signal strength is assumed from the signal characteristics and noise characteristics, and the probability of signal occurrence based on the distribution of many signal acquisition results is maximized. Estimate the true signal strength. For this reason, for example, a result that is not correctly measured in the signal acquisition result due to an external factor such as the eye moving during measurement and the intensity of the reflected signal from the eye changing, or the amount of noise changing due to disturbance. If it is included, it leads to fluctuation of the estimation result.

  Here, when the measured signal strength is large with respect to the change in signal strength caused by an external factor, the influence is very small. However, when the measured signal strength is small, a minute change in signal strength due to an external factor may cause a large fluctuation in the estimation result. As a result, for example, when a tomographic image of the fundus is taken, the roughness may be conspicuous on the image, particularly in the vitreous body or sclera, which is a low signal intensity region.

  On the other hand, in the range MAP estimation process, not only the signal intensity of the target pixel but also the signal intensity of the surrounding pixels is used for the estimation process, so that the estimated value is smoothed as in the case of applying a filter such as a moving average filter. can do. For this reason, noise in the vitreous body region or the like can be reduced by performing the range MAP estimation processing.

  FIG. 27A shows an example of a high-contrast tomographic image generated by a normal MAP estimation process using N pieces of tomographic images. FIG. 27C shows a one-line profile of a high-contrast tomographic image corresponding to the alternate long and short dash line in FIG. In particular, when a one-line profile of a high-contrast tomographic image corresponding to an A-scan image is seen, it can be seen that noise appears with high contrast as shown in FIG.

  On the other hand, FIG. 27B shows an example of a high-contrast tomographic image generated by the range MAP estimation process using N pieces of tomographic images. FIG. 27D shows a one-line profile of a high-contrast tomographic image corresponding to the one-dot chain line in FIG. In particular, when a one-line profile of a high-contrast tomographic image corresponding to the A scan image is seen, it can be seen that noise is reduced as shown in FIG. As described above, by performing the range MAP estimation process, it is possible to reduce noise in the vitreous body region or the like.

  As described above, the control unit 2400 according to the present embodiment includes the image generation unit 2410. The image generation unit 2410 performs estimation of the estimated value at the pixel position by performing the range MAP estimation process using the information of the plurality of tomographic images corresponding to the pixel position of the tomographic image among the information of the plurality of tomographic images. A portion 2416 is provided. The estimation unit 2416 performs range MAP estimation processing using information of a plurality of tomographic images corresponding to pixel positions for estimating an estimated value and pixel positions around the pixel position, among the information of the plurality of tomographic images. Estimate the value.

  Regarding the range MAP estimation processing, specifically, the estimation unit 2416 uses a predetermined probability density function for information on a plurality of tomographic images corresponding to the same pixel position and pixel positions around the pixel position. A probability density function of information of a plurality of tomographic images corresponding to pixel positions is calculated. Thereafter, the estimation unit 2416 calculates a product of the probability density functions using the probability density function of the calculated information of the plurality of tomographic images. Then, the estimation unit 2416 determines information on a tomographic image having the maximum probability in the product of probability density functions, and uses the determined value of information on the tomographic image as an estimated value.

  As described above, the control unit 2400 according to the present embodiment obtains an estimated value by the range MAP estimation process, and generates a tomographic image using the estimated value as a representative value. In the range MAP estimation process, the signal strength at the pixel positions around the target pixel is used in addition to the signal strength of the target pixel as the signal strength used for estimation of the estimated value. Therefore, it is possible to generate a substantially equivalent high-contrast tomographic image using less tomographic information than in the past. Therefore, the time required for imaging can be shortened, and the burden on the subject can be reduced. Further, in the range MAP estimation process, not only the signal intensity of the target pixel but also the signal intensity of surrounding pixels is used for the estimation process, so that noise in the vitreous region or the like can be reduced.

  In this embodiment, the luminance value of the tomographic image is used as the signal intensity in the addition averaging process and the range MAP estimation process. However, the signal intensity used in these processes is not limited to the luminance value. Since the signal used for these processes may be a signal corresponding to each pixel position, it may be a tomographic signal based on a signal obtained by performing Fourier transform on the interference signal acquired by the OCT optical system. The tomographic signal includes, for example, a signal obtained by subjecting the interference signal to Fourier transform, a signal obtained by subjecting the signal after Fourier transform to arbitrary signal processing for image generation, and a signal such as a luminance value based thereon.

  In the present embodiment, an example in which the fundus oculi Er is photographed is shown. However, the photographing target portion is not limited to the fundus oculi Er, and may be an anterior eye segment.

(Modification of Example 7)
In Example 7, in the range MAP estimation process, the signal intensity of the target pixel and the signal intensity of any of the surrounding pixels are assumed to have the same value, and used for estimation of the estimated value. On the other hand, in the modified example of the seventh embodiment, the signal intensity of the target pixel and the surrounding pixels is assumed to have a different contribution to the calculation of the estimated value depending on the pixel position (coordinates) of the pixel, and the estimated value is estimated. The probability density function used in the above is weighted according to the coordinates.

Specifically, as estimator 2416 that obtains the signal strength a _n of coordinates corresponding to the target pixel and surrounding pixels by the weighting factor W content corresponding to a desired contribution in step S2605, corresponding to the signal strength a _n The probability density function to be used is used to calculate the product of W probability density functions. That is, if the weight coefficient W for a certain coordinate 2, as obtaining the signal strength a _n of the coordinates twice, the probability density function corresponding to the signal strength a _n when calculating the product of the probability density function Use twice each. Similarly, the probability density function each true signal strength ν ₁ ~ν _max probabilities were power by respective weight coefficients W content in accordance with the coordinates included in the probability density function corresponding to the signal strength a _n at a certain coordinate The product of the probability density function may be used. The weighting factor W can be arbitrarily set according to a desired configuration. For example, the weighting factor W may be set higher for the surrounding pixels in the horizontal direction or the oblique direction of the target area according to the scanning direction of the measurement light. In this modification, the weighting factor W is set high for surrounding pixels closer to the target area.

FIGS. 28A to 28C show examples of the weighting factor W in this modification. FIG. 28A shows an example of the weighting factor W when the upper, lower, left, and right pixels of the target pixel are the surrounding pixels, as in the seventh embodiment. Here, for example, the probability density function of the signal strength _{a n} in a subject pixel is _{P t,} the probability density function of the signal strength _{a n} in the peripheral pixel is _P p1 _{to P p4.} In this case, the estimation unit 2416 uses the probability density function P _t related to the target pixel twice and calculates the probability density functions P _{p1 to} P _p4 related to surrounding pixels only once to calculate the product of the above-described probability density functions. The estimation unit 2416 uses a probability density function obtained by squaring the probabilities of the true signal intensities ν _{1 to} ν _max included in the probability density function P _t to calculate the product of the probability density functions. P _{p1 to} P _p4 may be used as they are (as a probability density function obtained by raising each probability to the first power). Since N probability density functions are calculated for each pixel, the weighting factor W is applied to each of the N probability density functions calculated for each pixel. Thereby, in the calculation of the product of the probability density function, the contribution degree of the probability density function of the target pixel can be made higher than that of the surrounding pixels, and the true signal intensity ν of the target pixel can be estimated more accurately.

FIG. 28B shows an example of the weighting factor W when the surrounding pixels are 8 pixels around the target pixel. Similarly, FIG. 28C shows an example of the weighting coefficient W when the surrounding pixels are 24 pixels around the target pixel. In these examples, the peripheral pixels closer to the target pixel can have a higher contribution to the calculation of the product of the probability density function, and the seventh embodiment corresponds to a case where a uniform weight coefficient is applied to the target pixel and the peripheral pixels. In comparison, the true signal intensity ν of the target pixel can be estimated more accurately. As described above, the weighting factor may be arbitrarily set according to a desired configuration. For example, it may follow a Gaussian distribution or another distribution. Note that the weighting method is not limited to the above method, and the probability of each of the true signal intensities ν _{1 to} ν _max in the product of the probability density function for obtaining the maximum probability in the posterior probability depends on the pixel position. Any method that reflects the weight may be used.

  As described above, in the OCT apparatus according to the present modification, the estimation unit 2416 weights the probability density function of information of a plurality of tomographic images corresponding to the pixel position for which the estimated value is obtained and the pixel positions around the pixel position. The product of the weighted probability density function is calculated. More specifically, the estimation unit 2416 performs weighting so that the probability density function of the information of a plurality of tomographic images corresponding to the pixel positions closer to the pixel position for which the product of the probability density functions is obtained becomes heavier. Do. As a result, the true signal intensity ν of the target pixel can be estimated more accurately.

(Example 8)
In the OCT apparatus according to Example 7, a high-contrast tomographic image was obtained using a small amount of tomographic information by using surrounding pixels. However, since the estimated value is estimated from the product of the probability density distribution including the probability density distribution of the surrounding pixels and one representative value is determined, it is generated compared to the case where the probability density distribution of only the target pixel is used. The resolution of the tomographic image is reduced. Therefore, in the OCT apparatus according to the eighth embodiment, a range MAP estimation process that uses surrounding pixels in a region without a fine (fine) structure that does not essentially require high resolution, and only a target pixel is used in other regions. A tomographic image is generated using a conventional MAP estimation process to be used.

  When obtaining a tomographic image of the fundus, the tomographic structure of the retinal tissue is usually observed, but at the same time, the vitreous body existing inside the eyeball may be observed. This is because various effects such as adverse effects on the retina due to the retina being pulled by the vitreous body are considered. The vitreous is made of a transparent material that allows light to enter the eye. For this reason, it is well known that when the vitreous body is observed by OCT using light, the substance is transparent and can only be observed with a very weak signal intensity. It is also known that the vitreous body does not have a fine structure compared to the retina.

  In addition, lesions are usually observed in the retina and choroid, and these films are observed with higher signal intensity in OCT than in the vitreous body. Therefore, by performing the range MAP estimation processing only in a region such as a vitreous body having a signal intensity lower than that of the retina, the contrast of the vitreous body or the like can be increased without degrading the resolution of a lesion appearing in the retina or the like.

  Hereinafter, the OCT apparatus according to the eighth embodiment will be described with reference to FIGS. In the OCT apparatus according to the present embodiment, the vitreous body region is specified as the target region. Therefore, the OCT apparatus according to the present embodiment applies the range MAP estimation process only to the vitreous region having no fine structure and low signal intensity, and performs the conventional MAP estimation process on the retina. Constituent elements of the OCT apparatus according to the present embodiment are the same as those of the OCT apparatus according to the seventh embodiment, and description thereof is omitted using the same reference numerals. Hereinafter, the OCT apparatus according to the present embodiment will be described focusing on differences from the OCT apparatus 1 according to the seventh embodiment.

  FIG. 29 shows a schematic configuration of the control unit 2900 according to the present embodiment. The image generation unit 2910 of the control unit 2900 includes an estimation unit 2916 and a region specifying unit 2917 in addition to the anterior eye image generation unit 211 to the calculation unit 215.

  The estimation unit 2916 performs normal MAP estimation processing or range MAP estimation processing on the plurality of tomographic images generated by the tomographic image generation unit 213, and estimates an estimated value of the true value of the interference signal. Note that the estimation unit 2916 may estimate the estimated value at each pixel position based on the signal after Fourier transform acquired by the acquisition unit 220 or the like.

  Based on the tomographic image generated by the tomographic image generation unit 213, the region specifying unit 2917 specifies a target region having a low signal intensity to which the range MAP estimation process should be applied. Note that the region specifying unit 2917 may specify the target region based on the Fourier-transformed signal acquired by the acquiring unit 220 or the like.

  Note that the estimation unit 2916 and the region specifying unit 2917 can also be configured by modules executed by the CPU or MPU of the control unit 2900. Further, the estimation unit 2916 and the region specifying unit 2917 may be configured by a circuit or the like that realizes a specific function such as an ASIC.

  Next, photographing processing according to the present embodiment will be described with reference to FIG. FIG. 30 is a flowchart of tomographic image capturing processing according to the present embodiment. In the imaging process according to the present embodiment, a target area to which the range MAP estimation process is applied is specified, and the range MAP estimation process is applied to the specified target area, and the normal MAP estimation process is applied to other areas. This is the same as the photographing process according to the seventh embodiment. Therefore, the description is omitted except for the difference.

  If the tomographic image generation unit 213 determines in step S506 that the high contrast mode is ON, the process proceeds to step S3001.

  In step S3001, the region specifying unit 2917 performs segmentation processing on at least one of the N tomographic images that have been aligned, and specifies a vitreous region that has a low signal intensity as a target region. The specific process is the same as the process performed by the region specifying unit 214 according to the first embodiment, and a description thereof will be omitted.

  In this embodiment, the region specifying unit 2917 extracts an inner limiting membrane (ILM) by segmentation processing. FIG. 31 shows each region when the vitreous region is specified by the segmentation process. The region specifying unit 2917 determines that the vitreous region 3101 is the side opposite to the RPE boundary line 3103 (pupil side) with the ILM boundary 3102 as the boundary, and specifies the region on the pupil side from the ILM boundary 3102 as the vitreous region 3101. . Thereby, the region specifying unit 2917 specifies the vitreous region as a target region to which the range MAP estimation process is applied.

  The segmentation process may be performed using any known method other than the above method. Furthermore, in this embodiment, the vitreous region is the target region, but an arbitrary region such as a sclera region known as a low signal intensity region may be specified as the target region. In this case, the region specifying unit 2917 can specify a layer such as the sclera by the segmentation process, and can set the layer as a target region. The target region may be a plurality of regions such as a vitreous region and a sclera region.

  When the target area is specified, in step S3002, the estimation unit 2916 applies range MAP estimation processing to the target area. Here, the range MAP estimation process in the present embodiment is the same as the range MAP estimation process shown in FIG. 26, but since the target to be processed is a pixel in the target area, the estimation unit 2916 performs initial processing in step S2601. The position is determined as a pixel position in the target area. The initial position may be an arbitrary position as long as it is a position within the target region. For example, the estimation unit 2916 sets the coordinates having the smallest x and z values in the target region as the initial position.

  In step S2607, the tomographic image generation unit 213 determines the estimated value estimated by the range MAP estimation process by the estimation unit 2916 as a representative value at the pixel position corresponding to the coordinates (x, z), and the pixel value To adopt. In step S2608, the estimation unit 2916 determines whether or not processing has been performed for all pixel positions in the target region. If there is an unprocessed pixel position, the pixel position is not processed in step S2609. To the pixel position. It should be noted that surrounding pixels may protrude outside the target region at the peripheral portion of the target region. In this case, if the protruding pixel is a pixel in the image, the signal intensity of the pixel can also be used for estimation. On the other hand, since regions other than the target region are regions with high signal strength, if the signal strength for the pixels in the region is used for the range MAP estimation process of the target region with low signal strength, the estimated value is true signal strength ν. The value may be far from. Therefore, the number of pixels used for the range MAP estimation process may be reduced by the amount of the protruding pixels.

  Next, in step S3003, the estimation unit 2916 applies normal MAP estimation processing to an area other than the target area. Here, the normal MAP estimation process is the same as the MAP estimation process according to the fourth embodiment illustrated in FIG.

  Note that the estimation unit 2916 applies normal MAP estimation processing to regions other than the target region. Therefore, in this case, the estimation unit 2916 determines the initial position as a pixel position in an area other than the target area in step S1701 of the normal MAP estimation process shown in FIG. The initial position may be an arbitrary position as long as it is a position in an area other than the target area. For example, the estimation unit 2916 sets the coordinates having the smallest values of x and z among the regions other than the target region as the initial position. In step S1702, the estimated value estimated by the normal MAP estimation process by the estimation unit 2916 is determined as the representative value at the pixel position corresponding to the coordinates (x, z), and is adopted as the pixel value. In step S1703, the estimation unit 2916 determines whether or not processing has been performed for all pixel positions in the area other than the target area. If there is an unprocessed pixel position, in step S1704, the pixel position Is moved to an unprocessed pixel position.

  As described above, the image generation unit 2910 according to the present embodiment uses the information of a plurality of tomographic images corresponding to the pixel position where the estimated value is estimated and the pixel positions around the pixel position to estimate the estimated value. An area specifying unit 2917 for specifying an area is provided. More specifically, the region specifying unit 2917 specifies an anatomical layer (layer boundary) of the fundus based on at least one tomographic image among the plurality of tomographic images. Thereafter, the area specifying unit 2917 specifies the target area from the specified layer and eye structure. The tomographic image generation unit 213 performs a range MAP estimation process in the target region, and performs a normal MAP estimation process in a region outside the target region. More specifically, in the target region, the estimation unit 2916 performs range MAP estimation processing using information on tomographic images corresponding to the pixel position where the estimated value is estimated and the surrounding pixel positions, and estimates the estimated value. . On the other hand, in an area other than the target area, normal MAP estimation processing is performed using information of tomographic images corresponding only to the pixel position where the estimated value is estimated, and the estimated value is estimated.

  As a result, an image having a high resolving power is generated by a conventional MAP estimation process in an area other than the target area, such as the retina. For this reason, it is possible to generate a high-contrast tomographic image while maintaining a sufficient resolving power.

  Further, the region specifying unit 2917 specifies a vitreous region as a target region. Compared with a vitreous body having a low signal intensity, the signal intensity obtained in the retina is high, so the influence of noise is relatively small. Therefore, the retinal region does not require as many samples as the vitreous region, and thus a high-quality, high-contrast tomographic image of the retina can be obtained by performing normal MAP estimation processing. Note that the same effect can be obtained by specifying the sclera region as the target region.

  Furthermore, since the signal used for the normal MAP estimation process may be a signal corresponding to each pixel position, it may be a tomographic signal based on a signal obtained by performing Fourier transform on the interference signal acquired by the OCT optical system.

  Further, similarly to the first embodiment, the image generation unit 2910 uses a calculation value (addition average value, median value, mode value, or maximum value) related to the average tomographic image as a representative value in a region other than the target region. The value may be determined. Furthermore, the area specifying unit 2917 may specify the target area using the signal intensity for each pixel, as in the second embodiment.

  In the first to third embodiments, the MAP estimation process is performed as the process for estimating the true signal strength. However, the estimation process is not limited to this. The processing performed in the estimation processing is processing for estimating the true signal intensity by performing statistical processing on a plurality of tomographic signals, and in particular, processing for estimating the true value of the tomographic signal at the pixel position using a predetermined model distribution. If it is. In addition, when performing an estimation process including surrounding pixels, an estimated value is obtained by using a predetermined model distribution for a plurality of tomographic signals corresponding to the pixel position where the estimated value is estimated and the surrounding pixel positions. Any processing that estimates the true value of the tomographic signal at the pixel position to be estimated may be used.

  In the first to third embodiments, the acquisition unit 220 acquires the interference signal acquired by the imaging optical system 100 and the tomographic signal generated by the tomographic image generation unit 213. However, the configuration in which the acquisition unit 220 acquires these signals is not limited to this. For example, the acquisition unit 220 may acquire these signals from a server or an imaging device connected to the control unit 200 via a LAN, WAN, or the Internet.

  Further, in Embodiments 1 to 3, the configuration of the Michelson interferometer is used as the interference optical system of the OCT apparatus, but the configuration of the interference optical system is not limited to this. For example, the interference optical system of the OCT apparatus may have a Mach-Zehnder interferometer configuration. Furthermore, the configuration of the imaging optical system 100 is not limited to the above configuration, and a part of the configuration included in the imaging optical system 100 may be configured separately from the imaging optical system 100.

  In the imaging optical system 100, a fiber optical system using an optical coupler is used as a dividing unit. However, a spatial optical system using a collimator and a beam splitter may be used. Furthermore, although a dichroic mirror is used as an optical member that divides light for each wavelength, the optical member is not limited to this. For example, the light may be divided for each wavelength by using a mirror composed of a perforated mirror or a prism on which a hollow mirror is deposited.

  In the first to third embodiments, a spectral domain OCT (SD-OCT) apparatus using an SLD as a light source has been described as the OCT apparatus. However, the configuration of the OCT apparatus according to the present invention is not limited to this. For example, the present invention can be applied to any other type of OCT apparatus such as a wavelength sweep type OCT (SS-OCT) apparatus using a wavelength swept light source capable of sweeping the wavelength of emitted light.

  In the first to third embodiments, the eye to be examined is described as the subject. However, the subject is not limited to the subject eye, and may be, for example, skin or an organ. At this time, the present invention can be applied to medical equipment such as an endoscope in addition to the ophthalmologic apparatus. In this case, the segmentation process can be performed according to the structure of the subject.

(Other examples)
The present invention supplies a program that realizes one or more functions of the above-described embodiments to a system or apparatus via a network or a storage medium, and one or more processors in a computer of the system or apparatus read and execute the program This process can be realized. It can also be realized by a circuit (for example, ASIC) that realizes one or more functions.

  As mentioned above, although this invention was demonstrated with reference to embodiment and an Example, this invention is not limited to the said embodiment and Example. Inventions modified within the scope not departing from the spirit of the present invention and inventions equivalent to the present invention are also included in the present invention. Moreover, each above-mentioned Example and modification can be combined suitably in the range which is not contrary to the meaning of this invention.

1: OCT apparatus (optical coherence tomography apparatus), 100: imaging optical system, 200: control unit (image processing apparatus), 213: tomographic image generation unit (generation unit), 214: region specifying unit, 215: calculation unit, 216: estimation unit, 220: acquisition unit

Claims (18)

An image processing apparatus that processes tomographic images generated using a plurality of sets of tomographic signals acquired by performing multiple optical coherence tomographic imaging using measurement light controlled to scan substantially the same part of a subject. There,
An acquisition unit for acquiring the plurality of sets of tomographic signals;
Using the plurality of sets of tomographic signals, generating a representative value at each pixel position of the tomographic image, and generating the tomographic image using the representative value at each pixel position as a pixel value;
With
The generator is
Using at least one set of tomographic signals among the plurality of sets of tomographic signals, an area specifying unit for specifying a target area in which contrast decreases in the tomographic image;
An estimation unit that estimates an estimated value by performing statistical processing on a plurality of tomographic signals corresponding to the pixel position among the plurality of sets of tomographic signals at each pixel position of the target region;
Including
The image processing apparatus generates the tomographic image using the estimated value as the representative value in the target region.   The estimation unit estimates a true value of the tomographic signal at the pixel position using the distribution of a predetermined model for the plurality of tomographic signals corresponding to the pixel positions of the target region as the estimated value. The image processing apparatus according to claim 1.   3. The estimation unit according to claim 1, wherein the estimation unit estimates the estimated value at the pixel position by performing a MAP estimation process on the plurality of tomographic signals corresponding to the pixel positions of the target region. Image processing apparatus. The estimation unit includes
For a plurality of tomographic signals corresponding to each pixel position of the target region, using a predetermined probability density function, a probability density function of the plurality of tomographic signals corresponding to the pixel position is calculated,
Calculating a product of the probability density functions for the plurality of tomographic signals corresponding to the same pixel position;
Determining the value of the tomographic signal that has the highest probability in the product of the probability density function;
The image processing apparatus according to claim 3, wherein the value of the determined tomographic signal is used as the estimated value. The generation unit further includes a calculation unit that calculates a calculation value at the pixel position using a plurality of tomographic signals corresponding to the pixel position of the tomographic image among the plurality of sets of tomographic signals,
The calculated value is one of an arithmetic average value, a median value, a mode value, and a maximum value,
5. The image processing apparatus according to claim 1, wherein the generation unit generates the tomographic image using the calculated value as the representative value in an area other than the target area of the tomographic image. 6. The subject is an eye to be examined;
The image processing apparatus according to claim 1, wherein the region specifying unit specifies the target region from at least one set of tomographic signals among the plurality of sets of tomographic signals and an eye structure.   The region specifying unit specifies a layer boundary of the fundus in the tomographic image using at least one set of the tomographic signals of the plurality of sets of tomographic signals, and determines the target region from the layer boundary and the eye structure. The image processing apparatus according to claim 6, which is specified.   The image processing apparatus according to claim 5, wherein the area specifying unit specifies an area of a pixel position having the calculated value lower than a threshold set using the calculated value calculated for each pixel position as the target area.   6. The device according to claim 1, wherein the region specifying unit specifies, as the target region, a region having a tomographic signal value lower than a threshold value set using one set of tomographic signals among the plurality of sets of tomographic signals. An image processing apparatus according to claim 1.   The image processing apparatus according to claim 1, further comprising a storage unit that stores the plurality of sets of tomographic signals subjected to gradation conversion into 8-bit information. An image processing apparatus that processes a tomographic image generated using a tomographic signal acquired by optical coherence tomographic imaging of a subject,
An acquisition unit for acquiring the tomographic signal;
A generating unit that generates the tomographic image using the tomographic signal;
With
The generator is
Using the tomographic signal, an area specifying unit for specifying a target area in which contrast decreases in the tomographic image;
An estimation unit that estimates an estimated value by performing statistical processing on the tomographic signal of the target region of the tomographic signal;
Including
The generation unit generates the tomographic image using the estimated value as a pixel value in the target region. An image processing apparatus that processes a tomographic image generated using a tomographic signal acquired by optical coherence tomographic imaging of a subject,
An acquisition unit for acquiring the tomographic signal;
A generating unit that generates the tomographic image using the tomographic signal;
With
The generator is
An estimation unit that estimates an estimated value by performing statistical processing on a tomographic signal corresponding to a pixel position of the tomographic image among the tomographic signals;
A filter unit that applies a noise removal filter to a tomographic image generated using the estimated value as a pixel value;
An image processing apparatus. An image processing apparatus that processes a tomographic image generated using a tomographic signal acquired by optical coherence tomographic imaging of a subject,
An acquisition unit for acquiring the tomographic signal;
A generating unit that generates the tomographic image using the tomographic signal;
With
The generation unit includes an estimation unit that estimates an estimated value by performing statistical processing on a tomographic signal corresponding to a pixel position of the tomographic image among the tomographic signals,
The estimation unit estimates the estimated value by performing statistical processing on a tomographic signal corresponding to a pixel position around the pixel position and a pixel position around the pixel position of the tomographic signal,
The image processing apparatus generates the tomographic image using the estimated value as a pixel value. An imaging optical system for imaging the subject with optical coherence tomography;
An image processing apparatus according to any one of claims 1 to 13,
An optical coherence tomographic imaging apparatus. An image processing method for processing a tomographic image generated using a tomographic signal acquired by optical coherence tomographic imaging of a subject,
Obtaining the tomographic signal;
Generating the tomographic image using the tomographic signal;
Including
The step of generating the tomographic image includes
Using the tomographic signal to identify a target region where contrast is reduced in the tomographic image;
Estimating the estimated value by performing statistical processing on the tomographic signal of the target area of the tomographic signal;
Including
In the step of generating the tomographic image, an image processing method of generating the tomographic image using the estimated value as a pixel value in the target region. An image processing method for processing a tomographic image generated using a tomographic signal acquired by optical coherence tomographic imaging of a subject,
Obtaining the tomographic signal;
Generating the tomographic image using the tomographic signal;
Including
The step of generating the tomographic image includes
A step of estimating an estimated value by performing statistical processing on a tomographic signal corresponding to a pixel position of the tomographic image among the tomographic signals;
Applying a noise removal filter to a tomographic image generated using the estimated value as a pixel value;
Including an image processing method. An image processing method for processing a tomographic image generated using a tomographic signal acquired by optical coherence tomographic imaging of a subject,
Obtaining the tomographic signal;
Generating the tomographic image using the tomographic signal;
Including
The step of generating the tomographic image includes a step of estimating an estimated value by performing statistical processing on a tomographic signal corresponding to a pixel position of the tomographic image among the tomographic signals,
In the step of estimating the estimated value, the estimated value is obtained by performing statistical processing on a tomographic signal corresponding to a pixel position for estimating the estimated value and a pixel position around the pixel position in the tomographic signal. Estimate
An image processing method for generating the tomographic image using the estimated value as a pixel value in the step of generating the tomographic image.   A program that, when executed by a processor, causes the processor to execute each step of the image processing method according to any one of claims 15 to 17.

Images