JP2018110715A - Image processing device and image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide techniques of generating high-quality reconstructed images at high speed from a plurality of transmission images.SOLUTION: When a transmission image contains first pixels and second pixels which are imaged at different times while continuously changing the positional relationship between a light source, an imaging subject and a detector, a reconstruction processing unit performs a projection or back projection operation between the transmission image and a reconstructed image, and when doing so, selects a pixel on the reconstructed image that corresponds to the first pixels on the basis of the positional relationship between the light source, the imaging subject and the detector at the time when the first pixels were imaged, and selects a pixel on the reconstructed image that corresponds to the second pixels on the basis of the positional relationship between the light source, the imaging subject and the detector at the time when the second pixels were imaged.SELECTED DRAWING: Figure 8

Description

  The present invention relates to a technique for reconstructing a three-dimensional image of an object from a plurality of transmission images.

As a technique for obtaining a three-dimensional image or a tomographic image inside an object, a technique called CT (Computed Tomography) scan or simply CT is known. X-ray CT using X-rays has been put into practical use in medical inspection apparatuses, industrial inspection apparatuses, and the like.

  In X-ray CT, an object is placed between a light source (X-ray source) and a detector arranged opposite to each other, and a plurality of transmission images (also called projection data) are taken by the detector while changing the X-ray projection direction. . Then, based on a plurality of transmission images having different projection directions, the inverse problem is solved by computer processing to reconstruct a three-dimensional image inside the object. In the conventional X-ray CT, the so-called Stop & Go method, in which the imaging of the projected image and the change of the projection direction (relative movement of the object and the detector) are alternately performed, has been mainstream. In order to shorten the number of apparatuses, an apparatus that employs a system (referred to as a continuous imaging system) that continuously captures a transmission image in each projection direction while relatively moving an object and a detector.

  CMOS sensors used for the detector include a global shutter type and a rolling shutter type. The global shutter method is a method in which exposure and data reading are performed at the same timing for all the lines of the CMOS sensor, and the rolling shutter method is a method in which the timing of exposure and data reading is slightly shifted for each line. Since the rolling shutter type sensor is less expensive than the global shutter type, a rolling shutter type sensor is often employed as a CT detector.

  However, as a weak point of the rolling shutter type sensor, there is a problem that image distortion occurs when shooting a moving object (generally called “rolling shutter distortion”). Therefore, when the above-described continuous imaging method is performed using a rolling shutter sensor, each transmitted image is distorted, which may significantly reduce the accuracy of reconstruction and the quality of the reconstruction result.

  Heretofore, several methods for correcting rolling shutter distortion have been proposed. For example, Patent Document 1 discloses a method of performing an inter-frame analysis of a moving image to acquire a motion vector generated when the camera moves, estimating a rolling shutter distortion component from the motion vector, and correcting the image. Yes.

  However, the method of Patent Document 1 has a heavy calculation load for calculating a motion vector and is difficult to perform real-time processing. Further, in the method of correcting based on image processing as disclosed in Patent Document 1, if there is a portion where motion vector estimation or rolling shutter distortion correction fails, there is a concern that an artifact may occur in the reconstruction result. . Therefore, the method of Patent Document 1 is not suitable for an inspection apparatus that requires high speed and an inspection apparatus that requires high reliability.

Japanese Patent No. 5531194

  The present invention has been made in view of the above circumstances, and an object thereof is to provide a technique for generating a high-quality reconstructed image from a plurality of transmission images at high speed.

  In order to achieve the above object, the image processing apparatus according to the present invention images the electromagnetic wave transmitted through the object a plurality of times by the detector while changing the positional relationship among the light source, the object and the detector. A transmission image acquisition unit that acquires data of a plurality of transmission images, and a reconstruction processing unit that performs a reconstruction process for generating a reconstruction image inside the object from the plurality of transmission images. The transmission image includes first and second pixels captured at different timings while continuously changing the positional relationship between the light source, the object, and the detector, When performing a projection or backprojection operation between the transmission image and the reconstructed image in the reconstructing process, the reconstruction processing unit performs the pixel on the reconstructed image corresponding to the first pixel, At the timing when the first pixel is imaged And selecting a pixel on the reconstructed image corresponding to the second pixel at a timing at which the second pixel is imaged, based on a positional relationship between the light source, the object, and the detector. The image processing apparatus is selected based on a positional relationship among a light source, the object, and the detector.

  In the conventional reconstruction process, all pixels on one transmission image are projected on the assumption that the positional relationship between the light source, the object, and the detector (hereinafter referred to as “projection direction”) is the same. (Projection from a point on the reconstructed image to a point on the transmission image) and back projection (projection from a point on the transmission image to a point on the reconstructed image) are performed. Therefore, “the transmission image includes the first pixel and the second pixel captured at different timings while continuously changing the positional relationship between the light source, the object, and the detector.” In the case of the conventional reconstruction process, an error occurs in the projection and backprojection processes.

  On the other hand, in the present invention, considering that the projection direction differs between the first pixel and the second pixel, the points on the transmission image are reconstructed based on the projection direction at the timing when each pixel is actually imaged. Corresponding points on the image. Therefore, for any pixel, the projection and backprojection processes in the reconstruction process can be performed with high accuracy, and a high-quality reconstruction result can be obtained. In addition, image processing for correcting distortion and pixel values of the transmission image is not necessary, and it is only necessary to correct the selection of corresponding points in projection and backprojection, so that the processing is fast. In addition, problems such as artifacts due to incomplete image processing (correction errors) do not occur.

  “The transmission image includes a first pixel and a second pixel captured at different timings while continuously changing the positional relationship between the light source, the object, and the detector” For example, it occurs when a transmission image is captured by a continuous imaging method using a rolling shutter type detector having different imaging timing for each line. In other words, the method of the present invention can be preferably applied to a CT apparatus that captures a transmission image by a continuous imaging method using a rolling shutter type detector.

  A storage unit that stores corresponding coordinate data in which a correspondence relationship between the coordinates of the pixels on the transmission image and the coordinates of the pixels on the reconstructed image is defined; and the reconstruction processing unit includes the reconstruction unit When a projection or backprojection operation is performed between the transmission image and the reconstructed image in processing, a pixel on the reconstructed image corresponding to a pixel on the transmission image is selected with reference to the corresponding coordinate data It is preferable to do. According to this configuration, it is not necessary to calculate the correspondence between each pixel on the transmission image and each pixel on the reconstructed image during the reconstructing process, so that the reconstructing process can be further speeded up. .

  The reconstruction process may be a process using a filtered back projection method. The reconstruction processing using the filtered back projection method has an advantage that a relatively high quality reconstruction result can be calculated at high speed. Note that the present invention can be preferably applied to a reconstruction algorithm other than the filtered backprojection method as long as the algorithm requires projection or backprojection.

  The present invention can be understood as an image processing apparatus having at least a part of the above configuration or function. The present invention can also be understood as a CT apparatus including an image processing apparatus, an inspection apparatus using a reconstructed three-dimensional image, and a diagnostic apparatus. The present invention also includes an image processing method, a CT apparatus control method, an inspection method, a diagnostic method, a program for causing a computer to execute these methods, or such a program, including at least a part of the above processing. Can also be regarded as a computer-readable recording medium on which non-temporary recording is performed. Each of the above configurations and processes can be combined with each other to constitute the present invention as long as there is no technical contradiction.

  According to the present invention, a high-quality reconstructed image can be generated at high speed from a plurality of transmission images.

FIG. 1 is a diagram schematically showing a hardware configuration of a CT system according to an embodiment of the present invention. 2A to 2D are diagrams illustrating an example of a movement method of the light source, the object, and the detector. FIG. 3 is a block diagram showing functions related to the reconstruction processing of the image processing apparatus. 4A is a diagram showing the basic principle of imaging, and FIG. 4B is a diagram showing reconstruction processing. FIG. 5 is a diagram showing reconstruction processing by a simple back projection method. FIG. 6 is a diagram showing the timing of exposure and data capture in a rolling shutter type CMOS sensor. 7A to 7C are diagrams illustrating continuous imaging using a rolling shutter type detector. FIG. 8A is a diagram illustrating a conventional reconstruction process, and FIG. 8B is a diagram illustrating a reconstruction process in consideration of imaging timing for each pixel. FIG. 9 is a diagram showing an example of the corresponding coordinate data. FIG. 10 is a flowchart of the image processing apparatus. 11A to 11G are diagrams showing results of effect verification by computer simulation.

  Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings. However, the description of each configuration described below should be changed as appropriate according to the configuration of the system to which the invention is applied and various conditions, and is intended to limit the scope of the present invention to the following description. Absent.

(System configuration)
FIG. 1 schematically shows a hardware configuration of a CT system 1 according to an embodiment of the present invention. The CT system 1 is a system that acquires data of a three-dimensional image and a tomographic image inside the object 102 using CT (Computed Tomography) technology. The CT system 1 can be applied to, for example, a medical inspection apparatus that visualizes the inside of a human being or an animal and is used for diagnosis or inspection, an industrial inspection apparatus that non-destructively inspects the internal structure of an industrial product, and the like.

The CT system 1 generally includes an imaging device 10, a control device 11, and an image processing device 12. The imaging device 10 is a device that captures a transmission image of the object 102. The control device 11 is a device that controls the operation of each unit of the imaging device 10. The image processing apparatus 12 is an apparatus that performs image processing and arithmetic processing on transmission image data (also referred to as projection data) acquired by the imaging apparatus 10. The imaging device 10, the control device 11, and the image processing device 12 may be configured as an integrated device, or may be configured as separate devices.

  The imaging device 10 includes a light source 100 and a detector 101 that are disposed to face each other. By projecting an electromagnetic wave 103 from the light source 100 onto the object 102 and imaging the electromagnetic wave 103 transmitted through the object 102 with the detector 101, a transmission image of the object 102 is obtained. In the present embodiment, a cone beam X-ray source is used as the light source 100, and a two-dimensional X-ray detector composed of a scintillator and a two-dimensional CMOS sensor is used as the detector 101. A configuration for obtaining a line transmission image is adopted. However, as the light source 100, a fan beam type or parallel beam type light source may be used. Further, not only the X-rays but also electromagnetic waves (for example, gamma rays, visible light, etc.) having a certain transparency to the object 102 may be used for imaging.

  Further, the imaging apparatus 10 includes a moving mechanism (not shown) for changing the positional relationship among the light source 100, the object 102, and the detector 101 (that is, the projection direction of the electromagnetic wave on the object 102). As the configuration of the moving mechanism, for example, a type in which the light source 100 and the detector 101 rotate around the fixed object 102 as shown in FIG. 2A, and the light source 100 and the detector 101 are fixed as shown in FIG. 2B. The type in which the object 102 rotates on the spot, the type in which the light source 100 and the detector 101 rotate in a plane perpendicular to the axis passing through the object 102 as shown in FIG. 2C, and the object 102 in FIG. There are various methods such as a fixed type in which the light source 100 and the detection unit 101 move in parallel and in opposite directions.

  The control device 11 is a device that performs overall control of the CT system 1, such as projection / stop of the light source 100, imaging by the detector 101, and change of the projection direction (drive of the moving mechanism).

  The image processing device 12 is a device that performs image processing and arithmetic processing on transmission image data acquired by the imaging device 10. For example, reconstruction processing for generating a reconstructed image (also referred to as three-dimensional image or volume data) of the object 102 from a plurality of transmission image data having different projection directions, processing for generating an arbitrary tomographic image from the reconstructed image, Various processes can be implemented in the image processing apparatus 12 according to the purpose and application, such as a process of extracting feature amounts and performing inspection / diagnosis using reconstructed image and tomographic image data.

  The image processing apparatus 12 can be configured by a computer including a processor (CPU), a memory, a storage device (such as a hard disk drive), an input device (such as a mouse, a keyboard, and a touch panel), an output device (such as a display device), and the like. . The functions of the image processing device 12 to be described later are realized by the processor executing a program stored in the storage device. However, some or all of the functions of the image processing apparatus 12 may be configured by ASIC, FPGA, or the like. Alternatively, some or all of the functions of the image processing apparatus 12 may be executed by another computer, a server on a network, or the like.

FIG. 3 is a block diagram illustrating functions related to the reconstruction process of the image processing apparatus 12. The image processing apparatus 12 includes a transmission image acquisition unit 30, a reconstruction processing unit 31, and a parameter storage unit 32. The transmission image acquisition unit 30 has a function of acquiring data of a plurality of transmission images having different projection directions (that is, original data used for reconstruction). The transmission image acquisition unit 30 may directly acquire transmission image data from the imaging device 10 or may read transmission image data captured in the past from a storage device or an external data storage. The reconstruction processing unit 31 has a function of performing a reconstruction process for generating a reconstructed image inside the object 102 from a plurality of transmission images. As reconstruction algorithms, simple back projection, filtered back projection (SI) (Simultaneous Reconstruction Technique), ART (Algebraic Reconstruction Technique), search method (gradient method) (Gradient Method), conjugate gradient method, steepest descent method)), etc., and any algorithm may be used. In the present embodiment, a filtered back projection method having an advantage that a relatively high-quality reconstruction result can be calculated at high speed is used. The parameter storage unit 32 is a function for storing various parameters (setting values, definition data, tables, etc.) used in the reconstruction process.

(Principle of imaging and reconstruction)
Before describing the reconstruction processing of the present embodiment, the basic principle of imaging and reconstruction in a general CT apparatus will be described with reference to FIGS. 4A and 4B. For ease of explanation, FIG. 4A shows an example in which a one-dimensional transmission image is captured twice using a parallel beam and detectors 101 of three pixels L0 to L1. In an actual device, a two-dimensional transmission image is taken using a cone beam, and the number (number of projections) of the image is several tens to several hundreds, and the number of pixels of the transmission image is several tens of thousands to several millions. The general principle is the same as that of FIG. 4A.

  In FIG. 4A, the numerical value shown inside the object 102 represents the X-ray attenuation coefficient distribution, and the broken line arrow indicates the transmission path (projection line) of the X-ray 103. First, at time t <b> 0, the light source 100 and the detector 101 are arranged above and below the object 102, the X-ray 103 is projected from the light source 100, and the first transmission image 40 is captured by the detector 101. When the X-ray 103 passes through the object 102, a part of the energy is absorbed and attenuated. Usually, since the X-ray attenuation coefficient distribution inside the object 102 is not uniform, a difference occurs in the attenuation factor of the X-ray 103 depending on the transmission path (the negative sign of the natural logarithm of the attenuation factor of the X-ray 103). (This corresponds to the integral value of the X-ray attenuation coefficient on the transmission path.) However, the value of the transmission image 40 is obtained by dividing the observed value by the observed value when no object is present and taking the natural logarithm of that value and taking the negative sign.

  At time t1, the light source 100 and the detector 101 are moved, the projection direction is changed by 90 degrees clockwise, the X-ray 103 is projected from the light source 100, and the second transmission image 41 is photographed. Since the transmission path of the X-ray 103 is different between the first transmission image 40 and the second transmission image 41, it can be seen that different values are observed.

  In this manner, after obtaining a plurality of transmission images 40 and 41 having different projection directions, reconstruction processing is performed using the transmission images 40 and 41. As shown in FIG. 4B, the reconstruction process corresponds to a process of estimating the X-ray attenuation coefficient distribution 42 inside the object 102 based on the plurality of transmission images 40 and 41.

  FIG. 5 schematically shows reconstruction processing by the simple back projection method. First, a data area of the reconstructed image 50 corresponding to one cross section of the object 102 is secured on a memory and initialized with zero. Then, the transmission image 40 is back-projected on the reconstructed image 50. Backprojection is an operation of projecting a point on the transmission image 40 onto a point on the reconstructed image 50. Here, the value of the pixel on the transmission image 40 is written to the corresponding pixel on the reconstructed image 50 (addition). Process). The corresponding pixels are pixels that exist on the X-ray transmission path. In the example of FIG. 5, three corresponding pixels exist for each pixel of the transmission image 40. The same value is written (added) to the three corresponding pixels. Thereafter, the transmission image 41 is also back-projected onto the reconstructed image 50 in the same manner. The reconstructed image 50 can be obtained by performing this process on all the transmission images and cumulatively adding back projections of all the transmission images. In FIG. 5, the back projection is cumulatively added and then divided by the number of projections (here, twice). The above is the outline of the reconstruction processing by the simple back projection method. The filter-corrected backprojection method used in this embodiment is characterized in that a high-pass filter is applied to the transmission image and then backprojected, and other operations are basically the same as the simple backprojection method.

(Continuous imaging system and its problems)
The imaging apparatus 10 according to the present embodiment uses a system (continuous imaging system) that continuously captures transmission images in each projection direction while relatively moving the object 102 and the detector 101. This is because the continuous imaging method can improve the throughput as compared with the so-called Stop & Go method in which the relative movement between the object 102 and the detector 101 is stopped to perform imaging.

  However, when the continuous imaging method is used, it is necessary to consider the influence of rolling shutter distortion generated by the detector 101. FIG. 6 is a diagram for explaining exposure / data fetching timing in a rolling shutter type CMOS sensor. The horizontal axis is time, and the vertical axis is the scan line of the CMOS sensor. It can be seen that the timing of exposure and data capture differs for each scan line. In the example of FIG. 6, the data acquisition of the scan line L1 is delayed by a time dt compared to the scan line L0, and the data acquisition of the scan line L2 is delayed by a time 2 × dt compared to the scan line L0. Due to this delay, image distortion called rolling shutter distortion occurs when a moving object is photographed. This distortion cannot be ignored as the relative speed between the object 102 and the detector 101 increases, and the accuracy of the reconstruction process and the quality of the reconstructed image are significantly reduced.

  7A to 7C and FIG. 8A will be used to describe in detail the problem when continuous imaging is performed using the rolling shutter type detector 101. FIG. In the case of the Stop & Go method (see FIG. 4A), data acquisition of all the scan lines L0 to L2 is performed in a state where the positional relationship among the light source 100, the object 102, and the detector 101 is not changed. On the other hand, in the case of the continuous imaging method, as shown in FIGS. 7A to 7C, the positional relationship (projection direction) among the light source 100, the object 102, and the detector 101 is different for each scan line. Therefore, although the first pixel 700, the second pixel 701, and the third pixel 702 are included in the same transmission image 70, the positional relationship between the light source 100, the object 102, and the detector 101 The data is captured with different (projection directions). Therefore, when the reconstruction process is performed using the transmission image 70, an error occurs if the back projection directions of all the pixels 700 to 702 are uniform as shown in FIG. 8A.

(Reconfiguration process)
In view of the above problems, in this embodiment, the influence of rolling shutter distortion is determined by determining the corresponding pixel at the time of back projection in consideration of the actual imaging timing (data acquisition timing) for each pixel and the projection direction at that time. To avoid. FIG. 8B schematically shows back projection in the reconstruction processing of the present embodiment. As shown in FIG. 8B, when the first pixel 700 in the transmission image 70 is back-projected, the corresponding pixel on the reconstructed image 80 is based on the projection direction at the timing t00 when the first pixel 700 was captured. (Gray pixel) is selected. When the second pixel 701 is back-projected, the corresponding pixel (gray pixel) on the reconstructed image 80 is selected based on the projection direction at the timing t01 when the second pixel 701 is imaged. When the third pixel 702 is back-projected, the corresponding pixel (gray pixel) on the reconstructed image 80 is selected based on the projection direction at the timing t02 when the third pixel 702 was captured. By adopting such a method, the accuracy of the reconstruction process can be improved.

  The correspondence between the pixels on the transmission image and the pixels on the reconstructed image is determined by the arrangement of the light source 100, the object 102, and the detector 101, the specifications of the detector 101 (number of pixels, pixel size, imaging timing for each scan line (frame Rate)), the moving speed of the detector 101, and the like, if the physical specifications of the imaging apparatus 10 and the control conditions during continuous imaging are known, they can be obtained by geometric calculation. Therefore, in this embodiment, for all transmission images acquired during continuous imaging, the correspondence between the pixels on the transmission image and the pixels on the reconstructed image is calculated in advance, and the corresponding coordinate data defining the correspondence is stored as a parameter. Stored in the unit 32. FIG. 9 shows an example of the corresponding coordinate data. By referring to the corresponding coordinate data when performing a projection or backprojection operation in the reconstruction process, it is possible to speed up the process rather than calculating the corresponding pixels one by one.

  FIG. 10 shows a processing flow of the image processing apparatus of this embodiment. In step S100, the transmission image acquisition unit 30 acquires data of a plurality of transmission images having different projection directions. In step S <b> 101, the reconstruction processing unit 31 reads the corresponding coordinate data from the parameter storage unit 32. In step S102, the reconstruction processing unit 31 applies a high-pass filter to each transmission image, and then performs back projection to generate a reconstructed image. At the time of back projection, as shown in FIG. 8B, correspondence between points on the transmission image and points on the reconstructed image is taken based on the projection direction at the timing when each pixel is actually imaged. In step S103, the reconstruction processing unit 31 displays a three-dimensional image or tomographic image of the object 102 on the display device based on the reconstructed image.

  According to the reconstruction process of the present embodiment described above, the backprojection process can be accurately performed for any pixel on the transmission image, and a high-quality reconstruction result can be obtained. Further, image processing for correcting distortion of the transmitted image and pixel values is unnecessary, and the selection of corresponding points in back projection need only be corrected, so that the processing is fast. In addition, problems such as artifacts due to incomplete image processing (correction errors) do not occur.

  By using the reconstruction processing of this embodiment, a high-quality reconstruction result can be obtained even when a transmission image is captured by the continuous imaging method using the rolling shutter method detector 101. In other words, it is possible to increase the speed of the CT system while maintaining the same quality as the Stop & Go method (still imaging).

  Further, in this embodiment, by using the corresponding coordinate data, it is not necessary to calculate the correspondence between each pixel on the transmission image and the pixel on the reconstructed image in the reconstruction process. Further speedup can be achieved.

(Example)
FIG. 11A to FIG. 11G show the result of effect verification by computer simulation. FIG. 11A is a model of the target object. Here, a spherical model 110 having a cavity 111 inside is used. FIG. 11B shows a result of generating X-ray transmission images in 32 directions for the model 110 by simulation, reconstructing these 32 transmission images by CT, and slicing the reconstructed image (three-dimensional image) at the center of the sphere. FIG. 11C and FIG. 11D. FIG. 11B shows a reconstruction result using a transmission image by the Stop & Go method (still image pickup method), that is, a transmission image without rolling shutter distortion. 11C and 11D are reconstruction results using a transmission image by a continuous imaging method, that is, a transmission image including rolling shutter distortion, and FIG. 11C is an example in which the conventional reconstruction processing (FIG. 8A) is applied. These are examples in which the reconstruction processing (FIG. 8B) of this embodiment is applied. 11E to 11G are profiles of broken line positions in FIGS. 11B to 11D, respectively, the horizontal axis represents the x coordinate of the image, and the vertical axis represents the normalized source weak coefficient.

PSNR (peak signal-to-noise ratio) is an index for quantitatively evaluating the accuracy of the reconstruction result, and is defined as follows. The larger the PSNR value, the better the accuracy (close to the quality of still imaging).

PSNR = 20 × log ₁₀ (MAXI / √MSE) [dB]
MAXI = maximum value of reconstruction result of still imaging MSE = (Σ (reconstruction result of still imaging−reconstruction result of continuous imaging) ² ) / number of pixels

In the case of the conventional reconstruction process, as shown in FIGS. 11C and 11F, the quality of the reconstruction result is significantly degraded due to the influence of rolling shutter distortion. On the other hand, it has been confirmed that by applying the reconstruction processing of the present embodiment, a reconstruction result having the same quality as that in the case of still imaging can be obtained as shown in FIGS. 11D and 11G. Further, when the PSNR values are compared, it is 11.0966 [dB] when the conventional reconstruction process is applied, whereas it is greatly increased to 41.6434 [dB] when the reconstruction process of the present embodiment is applied. It was confirmed that it improved.

(Other embodiments)
The above description of the embodiments is merely illustrative of the present invention. The present invention is not limited to the specific form described above, and various modifications are possible within the scope of the technical idea.

  For example, although the filtered back projection method is used in the above embodiment, a simple back projection method, a successive approximation method (SIRT method, ART method), a search method (gradient method, conjugate gradient method, steepest descent method)), etc. The reconstruction algorithm may be used. In any algorithm, at least one of back projection from the transmission image to the reconstructed image and projection from the reconstructed image to the transmission image is performed. When performing the back projection or projection operation, the rolling shutter is selected by selecting corresponding points between the transmission image and the reconstructed image based on the projection direction at the timing when each pixel is actually imaged, as in the above embodiment. It is possible to obtain a high-quality reconstruction result that is not affected by distortion.

1: CT system, 10: imaging device, 11: control device, 12: image processing device 30: transmission image acquisition unit, 31: reconstruction processing unit, 32: parameter storage unit 40, 41, 70: transmission image, 50, 80: Reconstructed image 100: Light source 101: Detector 102: Object 103: Electromagnetic wave 110: Model 111: Cavity 700: First pixel 701: Second pixel 702: Third pixel

Claims (6)

Transmission image acquisition for acquiring data of a plurality of transmission images obtained by imaging the electromagnetic wave transmitted through the object a plurality of times with the detector while changing the positional relationship between the light source, the object and the detector. And
A reconstruction processing unit that performs a reconstruction process for generating a reconstructed image inside the object from the plurality of transmission images,
The transmission image includes a first pixel and a second pixel captured at different timings while continuously changing a positional relationship between the light source, the object, and the detector;
The reconstruction processing unit
When performing a projection or backprojection operation between the transmission image and the reconstructed image in the reconstruction process,
A pixel on the reconstructed image corresponding to the first pixel is selected based on a positional relationship between the light source, the object, and the detector at a timing when the first pixel is imaged;
A pixel on the reconstructed image corresponding to the second pixel is selected based on a positional relationship between the light source, the object, and the detector at a timing when the second pixel is imaged. An image processing apparatus. A storage unit that stores correspondence coordinate data in which a correspondence relationship between the coordinates of the pixels on the transmission image and the coordinates of the pixels on the reconstructed image is defined;
The reconstruction processing unit corresponds to a pixel on the transmission image with reference to the corresponding coordinate data when performing a projection or back projection operation between the transmission image and the reconstruction image in the reconstruction processing. The image processing apparatus according to claim 1, wherein a pixel on the reconstructed image to be selected is selected. The detector is a rolling shutter type detector with different imaging timing for each line,
The image processing apparatus according to claim 1, wherein the first pixel and the second pixel are pixels captured by different lines of the detector. The image processing apparatus according to claim 1, wherein the reconstruction process is a process using a filtered back projection method. An image processing method executed by a computer,
Acquiring data of a plurality of transmission images obtained by imaging the electromagnetic wave transmitted through the object a plurality of times with the detector while changing the positional relationship between the light source, the object and the detector;
Performing a reconstruction process for generating a reconstructed image inside the object from the plurality of transmission images, and
The transmission image includes a first pixel and a second pixel captured at different timings while continuously changing a positional relationship between the light source, the object, and the detector.
When performing a projection or backprojection operation between the transmission image and the reconstructed image in the reconstruction process,
A pixel on the reconstructed image corresponding to the first pixel is selected based on a positional relationship between the light source, the object, and the detector at a timing when the first pixel is imaged.
A pixel on the reconstructed image corresponding to the second pixel is selected based on a positional relationship between the light source, the object, and the detector at a timing when the second pixel is imaged. A featured image processing method.   A program for causing a computer to execute each step of the image processing method according to claim 5.

Images