KR20160117011A - Apparatus and Method for Detection of Regularized Layer of Visual Cells in Eyeball Tomography - Google Patents

Abstract

The present invention relates to an apparatus and a method for quantifying detection of a layer of visual cells in eyeball tomography. According to the present invention, the apparatus for quantifying detection of a layer of visual cells includes: an image input part inputted with a three-dimensional volume image of an eyeball, obtained by optical coherence tomography; an image pre-processing part for carrying out a pre-process by regularizing and down-sampling the signal strength of the three-dimensional volume image, removing noise from the three-dimensional volume image by using a three-dimensional anisotropic kernel mask, and reinforcing the edge; a visual cell extraction part differentiating the three-dimensional volume image and separately extracting IS, ILM, RPE visual cell layers based on the pole of the differentiated signal; and a visual cell quantification part inputted with the kind of eye lesions and quantifying each layer visual cell layer depending on the kind of eye lesions. The apparatus for quantifying detection of a layer of visual cells is capable of automatically detecting clinically significant visual cells and accurately detecting and quantifying the visual cell layers, which are deformed irregularly by ocular diseases, without errors.

Description

[0001] The present invention relates to an apparatus and method for quantifying a cytocell layer in an ocular tomographic image,

The present invention relates to a method and apparatus for quantifying and detecting a photoreceptor layer in a tomographic image of an eyeball.

Originally, the ocular cell layer is uniformly layered horizontally. In the case of diseased eye, the ocular cell layer is often irregularly deformed. Therefore, it is possible to judge the lesion of eye disease based on the shape and thickness of the photoreceptor layer.

In the past, the physician has clinically confirmed the shape of the photocell layer and judged the eye disease. However, there is a limit to the amount of time and manpower required for clinical diagnosis involving such an operation. In addition, various other tests such as fundus examination have been performed to diagnose eye diseases. However, there is a limitation in that it is difficult to detect minute initial changes due to diseases in the eye by these test methods alone.

Accordingly, apparatuses and methods for diagnosing eye disease symptoms in tomographic images obtained using a tomography apparatus such as optical coherence tomography (OCT) have been developed. For example, there has been proposed a method for diagnosing optic nerve disease by pattern analysis of retinal map in optical coherence tomography as in the prior art document.

Optical coherence tomography is a tool for non-invasively imaging 3-D tomographic images of eye and other human tissues using the difference in the time of arrival of light reflected by the light transmitted through the tissue. The optical coherence tomography system acquires tomographic images of the eyeball and successively accumulates tomographic images to acquire 3D tomographic images. Optical coherence tomography can be used to confirm the progression of the diseases occurring in the retina of the eye, since the 3D tomographic image of the eye tissue can be taken at a micrometer scale resolution.

However, the 3D high resolution image data obtained using the OCT and the OCT is used for the diagnosis of ocular disease, and the amount thereof is very large, so there is a need for automated analysis using a computer.

However, since the intraocular tissues with lesions are often irregularly deformed when compared with normal eyeballs, there are many errors in the automatic detection of the optic cell layer, which is an important measure for determining whether ophthalmic lesions are present There is a problem that the possibility of occurrence of an error may be increased even in automatically diagnosing an eye disease.

US Patent Publication No. US 2008-0309881 A1 "PATTERN ANALYSIS OF RETINAL MAPS FOR THE DIAGNOSIS OF OPTIC NERVE DISEASES BY OPTICAL COHERENCE TOMOGRAPHY" (Dec. 18, 2008).

The object of the present invention is to analyze a three-dimensional volume image obtained by photographing a tomographic defect of a eyeball, to automatically detect a clinically meaningable photoreceptor layer, and to detect an irregularly changed temporal cell layer The present invention provides a method for quantifying a cytoskeleton layer and an apparatus therefor.

According to an aspect of the present invention, there is provided an apparatus for detecting a photocell layer, comprising: an image input unit receiving an ocular tomography image; A photocell layer detector for detecting at least one photocell layer determined in advance in the ocular tomographic image; And a cytoskeleton quantification unit for inputting the ocular disease information and correcting the erroneously detected region in the detected ocular cell layer according to the input ocular disease information to quantify the ocular cell layer.

Here, the ocular tomography image may be an image photographed using an optical coherence tomography apparatus.

Here, the image input unit may receive the three-dimensional eyeball image or may receive the at least one two-dimensional eyeball image including the at least one two-dimensional eyeball image.

Here, the ocular tomographic image may be a layered structure of the photoreceptor layer including the macula of the retina.

Here, the image preprocessing unit for performing at least one of performing normalization of the signal size of the ocularly tomographic image, performing noise elimination filtering, or performing image sharpening signal processing and pre-processing the ocular tomographic image And the cytosolic layer detector detects the photocell layer in the pre-processed ocular tomographic image.

Here, the image preprocessing unit performs normalization for adjusting the signal size of the image according to a predetermined standard, downsampling to reduce the image size, median filtering and Gaussian filtering And performing image sharpening signal processing for enhancing the edge of the image. [0031] According to the present invention,

Here, the cytosolic layer detector sets an inspection line in a direction perpendicular to the layered structure of the photoreceptor layer in the ocular tomographic image, obtains an inspection line image signal component representing a change in the image signal size on the inspection line And analyzing the acquired image signal component on the inspection line to detect the photoreceptor layer.

Here, the cytosolic layer detecting unit may analyze the image signal component on the inspection line, and stack the detected cytoskeleton layer two-dimensionally or three-dimensionally to detect the photocell layer in two or three dimensions in the ocular tomographic image .

Here, the photodiode layer detector may divide the image signal component on the inspection line into an inner segment and an outer segment, and determine whether the image signal value of the inspection line image signal component is up or down And detecting at least one of the second and third photocell layers in the portion where the first photocell is to be detected.

Here, the photodetector layer detection unit detects one local minimum value located between two local maximum values and the two local maximum values in the inspection line image signal component, and detects the local minimum value of the inspection line image signal And the component is divided into the above-mentioned knife-cutting portion and the above-mentioned portion.

Here, the photocell layer detector may be configured to detect the first photocell layer with respect to a video signal component on the inspection line based on a point where the degree of increase of the video signal value is the maximum at the end portion, And the third photocell layer is detected based on a point at which the degree of descent of the image signal value is maximum at the outer periphery of the second photocell layer.

Wherein the cytosolic layer detector calculates a rise or a fall of a video signal value of the video signal component on the inspection line using a differential value for the video signal component on the inspection line.

Here, the first photocell layer may be an ILM (Inner Limiting Membrane) layer, the second photocell layer may be an inner segment (IS) layer, and the third photocell layer may be an RPE (Retinal Pigment Epithelium) layer.

Wherein the ocular disease information includes information on the type of ocular disease, and the cytological cell layer quantifying part measures at least one of the ocular cell layer selected in accordance with the ocular disease information in a two-dimensional or three-dimensional manner Detecting a region of the photocell layer in which the photocell layer is discontinuous or having a change in signal magnitude greater than or equal to a predetermined reference as an erroneous detection region and correcting the erroneous detection region in the photocell layer to quantify the photocell layer .

Here, when correcting the erroneous detection area in the photocell layer, the cytologic layer quantification part corrects the erroneous detection area using another photocell layer or uses the neighboring area of the erroneous detection area in the photocell layer, And interpolating the erroneous detection area.

Here, the cytological cell layer quantifying unit may measure the ILM layer and the IS layer of the photoreceptor layer detected by the photoreceptor layer detecting unit when the type of eye disease corresponds to a macular hole (MH) Wherein the ILM layer corresponding to a circular region is replaced by the IS layer and the type of eye disease is diabetic macular edema (DME) or branch retinal vein occlusion (BRVO) Wherein a discontinuous section having a change in the video signal size of a predetermined reference or more is detected for the IS layer or the RPE layer detected by the cytological cell layer detection section when it corresponds to one of Age-related Macular Degeneration (AMD) And interpolating the discontinuous section using the interpolation method when the discontinuous section is detected.

Herein, the eye disease information includes information on the type of eye disease, and the cytological cell layer quantification unit determines whether or not the position of the cytological cell layer is erroneously detected for at least one of the cytogenetic layers selected according to the eye disease information And correcting the position of the photoreceptor layer when the photoreceptor is erroneously detected as a result of the determination, thereby quantifying the photoreceptor layer.

Here, when the type of ocular disease corresponds to high myopia (HM), the quantification part of the cytological cell layer quantifies the area of at least one of the ILM layer, the IS layer and the RPE layer of the photoreceptor layer, (ERM) of the retinal membrane, and corrects the position of the photoreceptor layer in the section of the photoreceptor layer outside the retinal tomography image according to the determination result. The degree of rise of the image signal value at the end portion of the image signal component on the inspection line is calculated to detect two candidate layers in ascending order of the degree of increase, And correcting the candidate layer closer to the retinal macular region to the ILM layer.

According to another aspect of the present invention, there is provided an apparatus for detecting a photocell layer, comprising: an image input unit for receiving an ophthalmologic tomography image showing a layered structure of a photocell layer including an inner macula of a retina; The inspection line is set in a direction perpendicular to the layered structure of the photoreceptor layer in the ocular tomography image, an inspection line image signal component indicating a change in the image signal size on the inspection line is obtained, And a cytocentric layer detection unit for analyzing the image signal component and detecting at least one predetermined cytocentric layer determined in advance.

The apparatus may further include an image preprocessing unit for normalizing the signal size of the ocularly tomographic image, performing noise elimination filtering, and performing image sharpening signal processing to pre-process the ocular tomographic image, And detect the photocell layer in a tomographic image.

Here, the photodiode layer detector may divide the image signal component on the inspection line into an inner segment and an outer segment, and determine whether the image signal value of the inspection line image signal component is up or down And detecting at least one of the second and third photocell layers in the portion where the first photocell is to be detected.

Here, the photodetector layer detection unit detects one local minimum value located between two local maximum values and the two local maximum values in the inspection line image signal component, and detects the local minimum value of the inspection line image signal And the component is divided into the above-mentioned knife-cutting portion and the above-mentioned portion.

Here, the first photocell layer is an ILM (Inner Limiting Membrane) layer, the second photocell layer is an IS (Inner Segment) layer, and the third photocell layer is a RPE (Retinal Pigment Epithelium) layer. The first photocell layer is formed on the basis of a point at which the rising degree of the image signal value is maximum in the outer peripheral portion, with respect to the image signal component on the inspection line, The third photocell layer may be detected based on a point at which the degree of descent of the image signal value is maximum at the outer portion of the second photocell layer.

According to another aspect of the present invention, there is provided a method for detecting a photocell layer comprising: inputting an image of a tomography (Tomography) image; A photocell layer detecting step of detecting at least one predetermined photocell layer determined in the ocular tomographic image; And a cytocentric layer quantification step of inputting the ocular disease information and correcting the erroneously detected region in the detected ocular cell layer according to the input ocular disease information to quantify the ocular cell layer.

The method of detecting quantitation of a photoreceptor layer according to the present invention automatically detects a clinically significant specific eye cell layer (IS, ILM, RPE) of the eyeball, and is also robust against an irregularly distorted photoreceptor layer Layer is automatically quantified and detected.

1 is a block diagram of a cytoskeleton layer detection apparatus according to an embodiment of the present invention.
2 is a block diagram of a cytoskeleton layer detection apparatus according to another embodiment of the present invention.
FIG. 3 is a reference diagram for explaining a process of preprocessing an image of the ocular tomography.
4 is a detailed block diagram of the image preprocessing unit.
FIG. 5 is a reference diagram showing an oculomotor image as a set of a plurality of two-dimensional ocular tomographic images.
FIG. 6 is a reference diagram showing that the image signal component of the inspection line obtained for the ocularly tomographic image is normalized.
FIG. 7 is a reference diagram showing the result of analyzing image signal components on the inspection line.
8 is a reference view showing the first through third photocell layers detected from the image signal component on the inspection line.
9 is a result image obtained by detecting the photoreceptor layer in two dimensions in an ocular tomographic image.
10 is a detailed block diagram of the cytosolic layer detection unit.
11 is a reference diagram for explaining the operation of the ocular tomography image of the macular hole (MH) and the quantification part of the cytosolic layer thereof.
FIG. 12 is a reference diagram for explaining the operation of an ocular tomographic image of a diabetic macular edema (DME) or a branch retinal vein occlusion (BRVO) and the operation of a cytosolic layer quantification unit therefor.
FIG. 13 is an explanatory diagram for explaining the operation of the ocular tomography image of age-related macular degeneration (AMD) and the quantification of the cytoskeleton layer thereof.
FIG. 14 is a reference diagram for explaining the operation of an ocular tomography image of High Myopia (HM) and a cytoskeleton layer quantification unit therefor.
15 is a reference diagram for explaining the operation of an ocular tomography image of the epi-retinal membrane (ERM) and the cytosolic layer quantification unit therefor.
16 is a block diagram of a cytoskeleton layer detection apparatus according to another embodiment of the present invention.
17 is a flowchart of a method of detecting a photocell layer in accordance with another embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the drawings, the same reference numerals are used to designate the same or similar components throughout the drawings. In the following description of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear. In addition, the preferred embodiments of the present invention will be described below, but it is needless to say that the technical idea of the present invention is not limited thereto and can be variously modified by those skilled in the art.

The ocular cytoplasm layer is uniformly layered horizontally. In the case of diseased eye, the ocular cell layer is often irregularly deformed. Therefore, it is possible to judge the lesion of eye disease based on the shape and thickness of the photoreceptor layer. In order to acquire information on the morphology and thickness of the photoreceptor layer, a tomographic apparatus such as optical coherence tomography (OCT) can be used.

Optical coherence tomography is a tool for non-invasively imaging 3-D tomographic images of eye and other human tissues using the difference in the time of arrival of light reflected by the light transmitted through the tissue. The optical coherence tomography system acquires tomographic images of the eyeball and successively accumulates tomographic images to acquire 3D tomographic images. Optical coherence tomography can be used to confirm the progression of the diseases occurring in the retina of the eye, since the 3D tomographic image of the eye tissue can be taken at a micrometer scale resolution.

The three - dimensional high - resolution image data obtained by using the above - described coherent tomography and the coherence tomography are used for the diagnosis of ocular diseases, and the amount of the three - dimensional image is very large. Therefore, there is a need for automated analysis using a computer.

However, since the intraocular tissues with lesions are often irregularly deformed when compared with normal eyeballs, there are many errors in the automatic detection of the optic cell layer, which is an important measure for determining whether ophthalmic lesions are present There is a problem that the possibility of occurrence of an error may be increased even in automatically diagnosing an eye disease.

For example, when a disease occurs on the retina, the horizontally uniform layer of the photoreceptor layer becomes deformed. For example, in the case of central retinal pigment epithelial disease, a fluid layer is formed in a specific layer of the photoreceptor cell. In the case of a macular hole disease, a hole is formed in a specific layer of the photoreceptor cell Is a key feature. In other words, the optic cell layer of the eye with the above-mentioned retinal disease has a lesion in which an abnormal region is formed in a specific layer of the photocell in a certain distance from the macular lutea.

For reference, the macula refers to the part of the eye where the light enters vertically into the center of the cornea and lens, which is depressed by a certain distance in the center of the retina. Also, a lesion means a change in a living body caused by a disease.

Disclosure of Invention Technical Problem [8] To solve the above problems, the present invention discloses a cytoskeleton layer detection apparatus capable of detecting a cytoskeleton layer strongly against a cytoskeleton layer deformed according to an eye disease, and a method therefor. The present invention also discloses a method and apparatus for detecting a photoreceptor layer capable of detecting a photoreceptor layer without causing an error in a general eyeball. In particular, the present invention discloses a photocell layer detector that can be applied to general eyeballs, and secondly, corrects a false detection region in a distorted photocell layer against a diseased photocell layer and detects a quantified photocell layer A cytoskeleton layer quantification section for quantifying the cytoskeleton layer.

The apparatus for detecting intra-retinal lesion according to the present invention can automatically detect a photoreceptor layer using an ocular tomography image obtained using the above-described optical coherence tomography.

However, the apparatus and method for detecting a photocell layer according to the present invention can automatically detect and quantify a photocell layer in the same manner using an ocular tomography image obtained through various methods other than the optical coherence tomography method. Therefore, the apparatus and method for detecting a photocell layer according to the present invention are not limited to the optical coherence tomography method.

Hereinafter, an apparatus and method for detecting a photocell layer in accordance with the present invention will be described in detail.

1 is a block diagram of a cytoskeleton layer detection apparatus according to an embodiment of the present invention.

The apparatus for detecting a photocell layer according to an embodiment of the present invention may include an image input unit 100, a photocell layer detector 200, and a photocell layer quantifier 300.

The image input unit 100 receives the tomography (tomography) image.

The ocular cell layer detector 200 detects at least one predetermined ocular cell layer in the ocularly tomographic image.

The cytological cell layer quantification unit 300 receives the ocular disease information and corrects the erroneously detected region in the detected ocular cell layer according to the inputted ocular disease information to quantify the cytological cell layer.

2 is a block diagram of a cytoskeleton layer detection apparatus according to another embodiment of the present invention.

In the above embodiment, the photocell layer detecting apparatus may further include an image preprocessing unit 150.

The image preprocessing unit 150 may perform at least one of normalizing the signal size of the ocularly tomographic image, performing noise removal filtering, or performing image sharpening signal processing, do.

When the image preprocessing unit 150 is further included, the cytosolic layer detector 200 detects the photocell layer in the pre-processed ocular tomographic image.

Hereinafter, each configuration of the photocell layer detecting apparatus according to the present invention will be described in more detail.

First, the image input unit 100 and the ocular tomography image will be described.

Herein, the ocular tomography image may be an image obtained by using an optical coherence tomography apparatus. Also, as described above, the ocular tomography image may be acquired using various tomography methods other than the optical coherence tomography method, and is not limited to the optical coherence tomography image.

Here, the image input unit 100 may receive the three-dimensional image of the ocularly-taken tomographic image, or may receive the ocular tomographic image including at least one two-dimensional ocular tomographic image. That is, the image input unit 100 can receive an ocular tomographic image, which is a three-dimensional volume image. Or an eyeball tomography image may be input as data composed of at least one two-dimensional eye tomography image.

Wherein the ocular tomographic image is a layered structure of the photoreceptor layer including the macula of the retina. In addition, the two-dimensional ocular tomographic image is preferably a tomographic image in which the lens of the eye is directed to one side and the macula of the eye is positioned on the other side and the eye is photographed in the direction crossing the eyeball. For example, a two-dimensional ocular tomographic image may be a tomographic image taken in a direction transverse to the eye, with the lens of the eye pointing upward and the macula of the eye pointing downward. Also, the 3D tomographic image can be three-dimensional image volume data aligned in the same manner as described above.

Next, the image preprocessing unit 150 will be described in more detail.

In order to accurately detect and quantify the photoreceptor layer in the ocular tomography image by the photoreceptor layer detector 200, it is necessary to pre-process the ocular tomography image input from the image input unit 100 according to the photoreceptor layer detection. This is because the quality of the acquired image may be changed depending on the photographing environment or the state of the vitreous body in the eyeball in the tomography. Therefore, in order to improve object detection performance, the image preprocessing unit 150 preferably performs at least one of noise removal from the ocular tomography image, normalization of the signal size, and signal processing for sharpening the image. That is, the image preprocessing unit 150 may normalize the signal size of the ocularly tomographic image, perform noise removal filtering, and perform image sharpening signal processing to pre-process the ocularly tomographic image.

Here, the image preprocessing may be performed on the entire tomographic image, which is three-dimensional volume data, or may be performed on each two-dimensional tomographic image.

First, the image preprocessing unit 150 prepares the image in preparation for the case where the intensity of the signal is not uniform or is shifted to one side of the histogram in the ocular tomography image inputted by the image input unit 100 The range of the video signal size of the pixels that are currently being processed can be normalized. Here, the image preprocessing unit 150 may perform normalization by changing the signal size of each pixel in the ocularly tomographic image according to a predetermined reference. For example, when the maximum value and the minimum value of the pixels coincide with predetermined values The signal size of the pixels in the tomographic image can be changed using a linear function or a non-linear function.

In addition, the image preprocessing unit 150 can reduce the size of the ocular tomographic image through down sampling in order that the ocular cell layer detector 200 detects the ocular cell layer more rapidly in the ocular tomographic image.

In addition, the image preprocessing unit 150 can remove noise through filtering in an eye tomography image. In the ophthalmologic tomography images acquired by optical coherence tomography, there is inherent noise generated in the imaging. Such a noise significantly reduces the sharpness of the image and is a main factor for lowering the performance of the cytosolic layer detector 200. Accordingly, the image preprocessing unit 150 may perform noise filtering by performing median filtering or Gaussian filtering. It is needless to say that the image preprocessing unit 150 may use various types of known noise elimination filters.

Here, the intraretinal viscous layer interlayer structure shows a pattern formed parallel to the horizontal direction (hereinafter referred to as the x-axis direction on the horizontal direction to the visual cell layer, and the y-axis direction on the vertical direction) Therefore, it is preferable that the image preprocessing unit 150 performs filtering using a kernel mask having anisotropic in the horizontal direction, that is, the x-axis direction, on the photoreceptor layer using the above characteristics. If the ocular tomography image is three-dimensional volume data, the filtering may be performed using a three-dimensional kernel mask. Also, if the 2D tomographic image is performed on a two-dimensional eye tomographic image, a two-dimensional kernel mask may be used.

In addition, the image preprocessing unit 150 may perform signal processing for sharpening the ocular tomographic image. Here, the image preprocessing unit 150 may perform image sharpening signal processing using an unsharp mask, which is a linear or nonlinear filter that amplifies a high frequency component of a signal. Alternatively, the image preprocessing unit 150 may perform signal processing for sharpening an eye tomography image or enhancing edges in an image using various masks or kernels. In addition, the image preprocessing unit 150 may perform signal processing for sharpening an image using various known image edge enhancement techniques.

FIG. 3 is a reference diagram for explaining a process of preprocessing an oculography image by the image preprocessing unit 150. Referring to FIG. (B) is an image obtained by performing filtering to remove noise from the image preprocessing unit 150, and (c) shows an image preprocessing unit 150 Image sharpening signal processing.

The image preprocessing unit 150 includes a video signal size normalization unit 151, a downsampling unit 152, a noise reduction filtering unit 153, and a noise reduction filtering unit 154. The image signal size normalization unit 151 performs each operation to perform the normalization, downsampling, noise removal, An image sharpening unit 154, and an image sharpening unit 154. [0034]

4 is a detailed block diagram of the image preprocessing unit 150 including the above-described sub-configuration.

As described above, the image preprocessing unit 150 performs normalization for adjusting the signal size of the image according to a predetermined reference, downsamples the image to reduce the image size, performs median filtering And performing Gaussian filtering to perform image sharpening signal processing for removing noise or enhancing an edge of an image by performing at least one of Gaussian filtering and Gaussian filtering.

When the image preprocessing unit 150 is further included, the cytosolic layer detector 200 can detect the photocell layer in the pre-processed ocular tomography image.

Next, the operation of the cytosolic layer detector 200 will be described in more detail.

The ocular cell layer detector 200 detects at least one predetermined ocular cell layer in the ocularly tomographic image.

The optic nerve cell layer, that is, the optic cell layer, is composed of various kinds of layers. Here, the optic cell layer detector 200 detects the intraocular lesion indicating a disease of the retina in order to diagnose a retinal disease, .

Here, it is preferable that the photoreceptor layer is any one of an inner (Inner Segment) layer, an RPE (Retinal pigment epithelium) layer, and an ILM (Inner limiting membrane) layer in the macula of the retina. Since the IS layer, the RPE layer, and the ILM layer are photocell layers containing clinically significant information in judging intraocular diseases, the photocell layer detector 200 preferably detects at least one of the photocell layers .

Referring to FIG. 9, the ILM layer L1, the IS layer L2, and the RPE layer L3 can be identified.

Next, a method of setting the data to be processed and the inspection line for detecting the photoreceptor layer in detecting the photoreceptor layer in the ocular tomography image will be described in detail first.

Herein, the photoreceptor layer detector 200 acquires a two-dimensional ocular tomographic image showing a cross-section of the photoreceptor layer in which the macula of the retina is positioned at one end in the ocular tomographic image, Can be detected.

When the image input unit 100 receives an ocular tomographic image, which is a three-dimensional volume image, the ocular cell layer detector 200 acquires the two-dimensional ocular tomographic image on the cross section of the ocular tomographic image, Can be detected. If the image input unit 100 receives the ocular tomographic image as the data including at least one two-dimensional ocular tomographic image, the ocular cell layer detector 200 detects the two-dimensional ocular tomographic image included in the ocular tomographic image It is possible to detect the photoreceptor layer.

Here, the two-dimensional ocular tomographic image is preferably a tomographic image in which the lens of the eye is directed to one side and the macula of the eye is positioned on the other side and the eye is photographed in the direction crossing the eye. For example, a two-dimensional ocular tomographic image may be a tomographic image taken in a direction transverse to the eye, with the lens of the eye pointing upward and the macula of the eye pointing downward.

Or the cytosolic layer detector 200 may detect the photocell layer directly from the 3D tomographic image. That is, the photodetector layer detector 200 may detect the photodetector layer directly from the three-dimensional volume data without acquiring the data representing the two-dimensional section from the three-dimensional volume data of the ocular tomographic image.

FIG. 5 is a reference diagram showing an oculomotor image as a set of a plurality of two-dimensional ocular tomographic images.

In order to detect a photoreceptor layer in an ocular tomographic image or a 2-dimensional ocular tomographic image, which is three-dimensional volume data, the photoreceptor layer detector 200 may be a 3-dimensional volume data, An inspection line is set in an image or a two-dimensional ocular tomographic image, an image signal component on an inspection line is acquired, and a photoreceptor layer is detected in the obtained image component of the inspection line image.

Here, the cytosolic layer detector 200 sets an inspection line in a direction perpendicular to the layered structure of the photoreceptor layer in the ocularly tomographic image and acquires a video signal component on the inspection line indicative of a change in the video signal size on the inspection line And analyze the image signal component on the obtained inspection line to detect the photoreceptor layer. Here, the video signal component on the inspection line is a signal component indicating a change in the intensity of the video signal. The size of the image signal of each pixel on the inspection line set in a direction perpendicular to the photoreceptor layer.

FIG. 6 is a reference view showing the image signal component of the inspection line obtained for the ocularly tomographic image as shown in FIG.

That is, the cytosolic layer detector may directly acquire the image signal component on the inspection line in the 3D tomographic image, detect the photoreceptor layer on the image signal component on the inspection line, The image signal component on the inspection line may be acquired to detect the photoreceptor layer.

Here, the cytosolic layer detector 200 analyzes the image signal component on the inspection line, and stacks the detected cytoskeleton layer two-dimensionally or three-dimensionally to detect the photocell layer in two or three dimensions in the ocular tomographic image . When the photoreceptor layer is detected in the two-dimensional ocular tomographic image, the photoreceptor layer may be detected on the two-dimensional ocular tomographic image. Alternatively, when a three-dimensional photocell layer is detected in the three-dimensional OCT, the photocell layer may be detected in a three-dimensional OCT image.

Next, detailed operation of the cytosolic layer detector 200 will be described in detail.

The cytosolic layer detector 200 may detect the photocell layer by normalizing the video signal size of the video signal component on the inspection line and analyzing the normalized inspection line video signal component. When the image preprocessing unit 150 is included, the normalized distribution of the image signal may be changed again through a preprocessing process such as noise removal filtering in the image preprocessing unit 150, Normalization is performed again to adjust the distribution of the image signal size for the tomographic image.

FIG. 6 is a reference diagram showing that the image signal component of the inspection line image obtained with respect to the ocularly tomographic image is normalized between 0 and 1 through the above normalization.

Next, the photic cell layer detector 200 may divide the video signal component on the inspection line into an inner segment and an outer segment.

Here, the photocell layer detector 200 detects one local minimum value located between two local maximum values and the two local maximum values in the image signal component on the inspection line, And the video signal component can be divided into the end portion and the non-end portion.

Here, the two local maximum values and the local minimum values of the video signal components on the inspection line can be detected using the differential values of the video signal components on the inspection line.

FIG. 7 is a reference diagram showing divided end portions and separate portions based on the two local maximum values H1 and H2, the local minimum value L1, and the local minimum value L1. Wherein the two local maximum values H1 and H2 represent the area where the retina starts and the macula area, respectively.

Next, the photocell layer detector 200 detects the first photocell layer at the inner edge portion or the second photocell layer at the outer edge portion based on the rising or falling degree of the image signal value of the image signal component on the inspection line, , And the third photocell layer can be detected.

Here, the first photocell layer may be an ILM (Inner Limiting Membrane) layer, the second photocell layer may be an inner segment (IS) layer, and the third photocell layer may be an RPE (Retinal Pigment Epithelium) layer.

Here, the photic cell layer detector 200 may calculate the rising or falling degree of the video signal value of the video signal component on the inspection line using the differential value for the video signal component on the inspection line.

Here, the photocell layer detector 200 may detect the first photocell layer based on a point at which the degree of increase of the image signal value is the maximum at the end portion of the image signal component on the inspection line.

In addition, the photocell layer detector 200 may detect the second photocell layer based on a point at which the degree of increase of the image signal value is maximum in the outer portion.

In addition, the photocell layer detector 200 can detect the third photocell layer based on a point at which the degree of descent of the image signal value is maximum at the outer edge portion.

8 is a reference view showing the first through third photocell layers detected in the video signal component on the inspection line as described above.

As described above, the cytosolic layer detector 200 may stack the cytoskeleton layer detected in the image signal component on the inspection line two-dimensionally or three-dimensionally, and detect the photocell layer in two or three dimensions in the ocular tomographic image .

9 is a result image obtained by two-dimensionally stacking the photoreceptor layer detected by the photoreceptor layer detecting unit 200 on the image signal component on the inspection line and detecting the photoreceptor layer in two dimensions in the ocular tomography image. As shown in FIG. 9, the photocell layer detector 200 can detect the ILM layer L1 as the first photocell layer, the IS layer L2 as the second photocell layer, and the RPE layer L3 as the third photocell layer.

In addition, the photocell layer detector 200 corrects the photocell layer position detected so as not to have continuity due to noise in the process of laminating the photocell layer, which is detected in the image signal component on the inspection line, two-dimensionally or three- Calibration can be performed to maintain continuity. Here, the photocell layer detector 200 may correct the data of discontinuous points when the curve or curved surface of the photocell layer stacked in two or three dimensions is discontinuous, so that the continuity of the photocell layer is maintained. For example, the cytosolic layer detector 200 can perform signal processing for performing interpolation of a discontinuous point using normally detected cytosolic layer data. Here, the correction performed by the photocell layer detector 200 differs from the photocell layer detector 200 in that the photocell layer quantifier 300, which will be described in detail below, corrects the erroneously detected area. It is premised that the photoreceptor layer is corrected in the image. For this, the photocell layer detector 200 determines whether the curved or curved surface of the photocell layer stacked in two or three dimensions is discontinuous, and determines whether the predetermined reference, that is, the degree of change in the signal size, It is determined whether or not the signal is discontinuous due to the noise, and the correction can be performed only when it is determined that the signal is discontinuous due to the noise.

FIG. 10 is a detailed block diagram of a cytoskeletal layer detection unit 200 including a sub-structure for performing each operation of the cytoskeletal layer detection unit 200 described above. The visual field layer detector 200 normalizes the image, the image signal component on the inspection line is divided into an inner and outer edge portions, and the first through third photoreceptor layers are detected. (201), a cytoskeleton layer structure analyzing unit (202), a cytoskeleton layer extracting unit (203), and a correcting unit (204).

The cytosolic layer detector 200 may divide the image signal component on the inspection line into an end portion and an outer portion, and may perform the normalization operation and the correction operation as necessary in the detection of the first through third photocell layers. Accordingly, the cytosolic layer detector 200 includes a cytoskeleton layer structure analyzer 202 and a cytoskeleton layer extractor 203, and may not include the image normalizer 201 or the corrector 204 as needed .

Next, the cytoskeleton layer quantification unit 300 will be described in more detail.

As described above, in a normal case, it is possible to normally detect the first through third photocell layers through the configuration of the cytocell layer detector 200 described above. However, in the case of eyeballs in which the retinal tissue is deformed due to the disease, the photoreceptor layer detecting unit 200 may erroneously detect the photoreceptor layer in the retinal tomographic image. The photocell layer quantification unit 300 according to the present invention performs an operation of correcting the photocell layer region that is erroneously detected due to the disease as described above and quantifying the photocell layer so that the photocell layer is normally detected.

The cytological cell layer quantification unit 300 receives the ocular disease information and corrects the erroneously detected region in the detected ocular cell layer according to the inputted ocular disease information to quantify the cytological cell layer.

Here, the eye disease information may include information on the types of eye diseases. In addition, the eye disease information may include information on the lesion area together with the type of eye disease.

Here, the photocell layer quantification unit 300 can quantify the photocell layer using other methods depending on the type of ocular disease and the shape of the retinal photocell layer. 11 to 15 are reference views showing a modification of the photoreceptor layer according to the type of disease. 12 shows the diabetic macular edema (DME) or branch retinal vein occlusion (BRVO), FIG. 13 shows the age-related macular hole (Age- related macular degeneration (AMD), FIG. 14 is high myopia (HM), and FIG. 15 is an image showing an ocular tomography image obtained in each case of an epi-retinal membrane (ERM).

First, when the discontinuous section appears in the detected photoreceptor layer, the photoreceptor layer quantifying section 300 may modify the photoreceptor layer so that the discontinuous section is continuous using the continuously detected photoreceptor layer around the photoreceptor layer.

11 and 12, a macular hole (MH), a diabetic macular edema (DME), a retinal vein occlusion (BRVO), an age-related macular hole Macular Degeneration, AMD), it can be seen that any one or more of the cytoskeletal layers are discontinuous or appear as areas with a change in signal amplitude above a predetermined reference.

Here, the photocell layer quantification unit 300 may be configured such that, in the photocell layer detected two-dimensionally or three-dimensionally by the photocell layer detector 200 for at least one photocell layer selected according to the eye disease information, It is possible to detect an area which is discontinuous or has a change in the signal size larger than a predetermined reference as an erroneous detection area.

At this time, the cytosolic layer quantification unit 300 can detect the erroneous detection region with respect to at least any one of the ILM layer, the IS layer, and the RPE layer according to the ocular disease information.

Here, the cytosolic layer quantification unit 300 may detect a corresponding section as an erroneous detection region when the photocell layer changes to a predetermined reference or more. In this case, the cytosolic layer quantification unit 300 calculates a differential value with respect to the cytosolic layer, determines whether the photocell layer changes to a predetermined reference or more on the basis of the calculated differential value, Can be detected.

It is possible to quantify the cytoskeleton layer by correcting the detected erroneous detection region as in the cytoskeleton layer quantification section 300. [ At this time, in correcting the erroneous detection region in the photoreceptor layer, the photoreceptor layer quantifying unit 300 may correct the erroneous detection region using another photocell layer, i.e., another photocell layer not erroneously detected. Alternatively, the cytosolic layer quantification unit 300 may correct the erroneous detection region of the cytosolic layer by performing interpolation using the neighboring region of the erroneous detection region in the cytosolic layer, that is, the neighboring region that is not erroneously detected in the cytosolic layer have.

First, a case in which the type of eye disease corresponds to a macular hole (MH) is described.

11 is a tomographic image of the eye with macular hole disease.

At this time, the cytosolic layer quantification unit 300 quantizes the ILM layer corresponding to the circular region, which is a lesion region corresponding to the macular hole, with the IS Layer.

In the macular hole, a macular hole is formed in the macula of the retina, and some macular tissue including the ILM layer is partially lost. Therefore, when the photoreceptor layer detecting unit 200 detects the photoreceptor layer on an ocular tomographic image including the macular hole disease, the ILO layer is determined as a layer structure in which the hole regions are not continuous as shown in FIG. 11 (b) There is a problem that it is detected incorrectly.

The temporal layer quantification unit 300 corrects the ILM layer by the above-described method. That is, the cytosolic layer quantification unit 300 sets an area corresponding to the hole in the ocular tomographic image and corrects the erroneous detection in the area. As described above, since the ILM layer detected in the set circular region is an erroneously detected layer that does not actually exist, the ILM layer of the corresponding region is the same as the normally detected IS layer , And the ILM layer in the region is replaced with the IS layer as shown in Fig. 12 (c).

Next, the type of ocular disease is classified into any one of diabetic macular edema (DME), branch retinal vein occlusion (BRVO), or age-related macular degeneration (AMD) Consider the case. In this case, the cytosolic layer quantification unit 300 detects a discontinuity section having a change in the video signal size that is equal to or larger than a predetermined reference to the IS layer or the RPE layer detected by the cytosolic layer detection section 200, The discontinuous section can be interpolated using interpolation.

Hereinafter, the case of diabetic macular edema (DME) or branch retinal vein occlusion (BRVO) is described in detail.

12 is a tomographic image of a diabetic macular edema (DME) or branch retinal vein occlusion (BRVO).

In the cases of diabetic macular edema (DME) and retinal vein occlusion (BRVO), shadow from the upper part of the retina in the ophthalmologic examination image is caused by exudate and hemorrhage of the retina, The IS layer or the RPE layer is lost as shown in (a) of FIG. Accordingly, when the photoreceptor layer detecting unit 200 detects the photoreceptor layer for the eye having the two diseases, the exudate and the bleeding unit are erroneously detected as the photoreceptor layer instead of the lost IS layer and the RPE layer as shown in FIG. 12 (b).

The temporal layer quantification unit 300 detects a discontinuity section having a change in the video signal size of a predetermined reference or more with respect to the IS layer or the RPE layer detected by the temporal layer detection unit 200 as described above, The discontinuous section can be interpolated.

As described above, a shadow is generated in a region with exudates and bleeding, and the photoreceptor layer is erroneously detected due to the disappearance of the photoreceptor layer in the retinal tomography image. As a result, the photoreceptor layer quantifying unit 300 performs the differentiation on the detected photoreceptor layer And if the magnitude of the differential value is larger than a predetermined reference, it can be detected as a section in which a discontinuous section appears.

At this time, the cytosolic layer quantification unit 300 can correct the cytoskeletal layer by interpolating the discontinuity section that has been erroneously detected as described above by using the interpolation method using the neighboring region in the photoreceptor layer as described above. Here, the cytosolic layer quantification unit 300 can perform interpolation on the discontinuous section of the cytosolic layer using various interpolation functions. For example, the cytosolic layer quantification unit 300 may use a spline interpolation method using a spline interpolation function.

FIG. 12C is a reference diagram showing the IS layer and the RPE layer interpolated and corrected as described above.

Next, the case where the type of eye disease is Age-related Macular Degeneration (AMD) will be described in detail.

Fig. 13 is a tomographic image of the eye with age-related macular degeneration (AMD).

In the case of AMD (AMD), lesions such as subretinal fluid, fibrovascular wound, pigment epithelium detachment, and protein residue appear. In this case, a common phenomenon in the interlayer structure of the cytosol is that the IS layer and the RPE layer are severely damaged as shown in FIG. 13 (a). Therefore, in this case, the cytosolic layer detector 200 erroneously detects the photocell layer including the IS layer and the RPE layer as shown in FIG. 13 (b).

As described above, the cytosolic layer quantification unit 300 detects a discontinuity section having a change in the video signal size of a predetermined reference or more with respect to the IS layer or the RPE layer detected by the cytosolic layer detection unit 200, The discontinuity section can be interpolated.

At this time, the cytosolic layer quantification unit 300 can correct the photoreceptor layer by interpolating the discontinuity section that has been eroded and erroneously detected using the interpolation method using the region normally detected in the photoreceptor layer as described above. In this case, since the photoreceptor layer is severely damaged, it is preferable to perform the interpolation using the linear interpolation method.

Next, the photocell layer quantification unit 300 can correct the photocell layer when the photocell layer is erroneously detected.

Here, the photocell layer quantification unit 300 may determine whether or not the position of the photocell layer is erroneously detected for at least one photocell layer selected according to the eye disease information. If the result of the determination is false, the position of the photoreceptor layer may be corrected to quantify the photoreceptor layer.

Next, the case where the type of eye disease corresponds to high myopia (HM) will be described in detail.

14 is a tomographic image of the eye with High Myopia (HM).

Here, the cytosolic layer quantification unit 300 determines whether at least one of the ILM layer, the IS layer, and the RPE layer among the photocell layers has a section deviating from the retinal tomography image, It is possible to correct the location of the photoreceptor layer in the section of the photoreceptor layer outside the tomographic image.

In the case of high myopia (HM), the curvature of the retina has an abnormal shape. Thus, as shown in FIG. 14 (a), the retina does not completely enter the field of view (FOV) There are many cases. Therefore, when the photoreceptor layer detecting unit 200 detects the photoreceptor layer in an ocularly acquired tomographic image of the eye having high myopia, an error occurs in detecting the position of the photoreceptor layer as shown in FIG. 14 (b) do.

The photocell layer quantification unit 300 may determine whether the photocell layer has a section deviated from the retinal tomography image and correct the position of the photocell layer when the photocell layer has deviated from the photocell layer. For example, the cytosolic layer quantification unit 300 can process the position of the photoreceptor layer in the section of the photoreceptor layer outside the retinal tomography image as a boundary of the retinal tomography image, as shown in FIG. 14 (c).

Next, the case in which the type of eye disease corresponds to an epi-retinal membrane (ERM) will be described in detail.

15 is a tomographic image of an eye with an epi-retinal membrane (ERM).

Here, the cytosolic layer dangling unit 300 calculates a rise degree of an image signal value at an end portion of a video signal component on the inspection line, detects two candidate layers in ascending order of the degree of increase, The candidate layer closer to the retinal macular region in the candidate layer can be corrected to the ILM layer.

In the case of ERM, another film appears in front of the retina as shown in FIG. 15 (a). The retinal whole membrane thus formed is present in the upper part of the retina and often has a high reflectance, and the membrane may peel off from the retina surface depending on the progress of the symptoms. As a result, the photoreceptor layer detector 200 erroneously detects the entire retina layer as an ILM layer and erroneously detects the position of the ILM layer as shown in FIG. 15 (b).

The temporal layer quantification unit 300 calculates a rise degree of an image signal value at an end portion of the image signal component on the inspection line, detects two candidate layers in ascending order of the degree of increase, It is preferable to correct the candidate layer closer to the retinal macular region in the candidate layer to the ILM layer. Here, the degree of elevation can be calculated using the differential value of the video signal component on the inspection line. 15 (c) is a reference diagram showing the result of correcting the ILM layer in this way.

16 is a block diagram of a cytoskeleton layer detection apparatus according to another embodiment of the present invention.

The apparatus for detecting a cytoskeleton layer according to another embodiment of the present invention may include an image input unit 100 and a cytosolic layer detector 200. Here, the image input unit 100 and the cytosolic layer detector 200 may be implemented in the same manner as the image input unit 100 and the cytoskeletal layer detector 200 of the cytoskeleton layer detection apparatus according to the present invention described above with reference to FIGS. 1 to 15 Can operate. The overlapping portions will be omitted and briefly described.

The image input unit 100 receives an ocular tomography image showing a layered structure of the photoreceptor layer including the macula of the retina.

The cytosolic layer detector 200 sets an inspection line in a direction perpendicular to the layered structure of the photocell layer in the ocular tomography image and acquires an inspection line image signal component indicating a change in the image signal size on the inspection line And analyzes at least one of the at least one photocell layer determined by analyzing the image signal component on the obtained inspection line.

Herein, the cytological cell layer detection apparatus may further include an image preprocessing unit 150 as needed.

The image preprocessing unit 150 normalizes the signal size of the ocularly tomographic image, performs noise elimination filtering, and performs image sharpening signal processing to pre-process the ocularly tomographic image.

When the image preprocessing unit 150 is included, the cytosolic layer detector 200 detects the photocell layer in the pre-processed ocular tomographic image.

The cytosolic layer detector 200 divides the image signal component on the inspection line into an inner segment and an outer segment, and detects a rise or a fall of the image signal value of the image signal component on the inspection line, , It is possible to detect the first photoreceptor layer in the above-mentioned incision portion or to detect at least one or more of the second photoreceptor layer and the third photoreceptor layer in the above-mentioned portion.

Here, the first photocell layer may be an ILM (Inner Limiting Membrane) layer, the second photocell layer may be an inner segment (IS) layer, and the third photocell layer may be an RPE (Retinal Pigment Epithelium) layer.

Here, the photocell layer detector 200 detects one local minimum value located between two local maximum values and the two local maximum values in the image signal component on the inspection line, And the video signal component can be divided into the end portion and the non-end portion.

Here, the photocell layer detector 200 may be configured to detect the first photocell layer with respect to the image signal component on the inspection line based on a point at which the degree of increase of the image signal value is the maximum in the second portion, The third photocell layer can be detected on the basis of a point at which the rising degree of the image signal value is maximum based on a point where the degree of descent of the image signal value is maximum in the outer photocell layer.

17 is a flowchart of a method of detecting a photocell layer in accordance with another embodiment of the present invention.

The method of detecting a photocell layer according to another embodiment of the present invention may include an image input step (S100), a photocell layer detection step (S200), and a photocell layer quantification step (S300). In addition, it may further include an image preprocessing step (S150) as necessary. Here, the image input step S100, the image preprocessing step S150, the photocell layer detection step S200, and the photocell layer quantification step S300 may be performed by using the photocell layer detection device 100 according to the present invention, The image processor 100, the image preprocessing unit 150, the cytosolic layer detector 200, and the cytoskeleton quantification unit 300 may operate in the same manner. The overlapping portions will be omitted and briefly described.

The image input step S100 receives the tomography image.

The photoreceptor layer detecting step (S200) detects at least one photoreceptor layer predetermined in the ocular tomographic image.

The photoreceptor layer quantification step (S300) receives the ocular disease information and corrects the erroneously detected region in the detected ocular cell layer according to the inputted ocular disease information to quantify the photoreceptor layer.

It is to be understood that the present invention is not limited to these embodiments, and all elements constituting the embodiment of the present invention described above are described as being combined or operated in one operation. That is, within the scope of the present invention, all of the components may be selectively coupled to one or more of them.

In addition, although all of the components may be implemented as one independent hardware, some or all of the components may be selectively combined to perform a part or all of the functions in one or a plurality of hardware. As shown in FIG. In addition, such a computer program may be stored in a computer readable medium such as a USB memory, a CD disk, a flash memory, etc., and read and executed by a computer to implement an embodiment of the present invention. As the recording medium of the computer program, a magnetic recording medium, an optical recording medium, a carrier wave medium, and the like can be included.

Furthermore, all terms including technical or scientific terms have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs, unless otherwise defined in the Detailed Description. Commonly used terms, such as predefined terms, should be interpreted to be consistent with the contextual meanings of the related art, and are not to be construed as ideal or overly formal, unless expressly defined to the contrary.

It will be apparent to those skilled in the art that various modifications, substitutions and substitutions are possible, without departing from the scope and spirit of the invention as disclosed in the accompanying claims. will be. Therefore, the embodiments disclosed in the present invention and the accompanying drawings are intended to illustrate and not to limit the technical spirit of the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments and the accompanying drawings . The scope of protection of the present invention should be construed according to the following claims, and all technical ideas within the scope of equivalents should be construed as falling within the scope of the present invention.

100:
150: Image preprocessing section
151: Video signal size normalization unit
152: Downsampling unit
153: Noise elimination filtering unit
154: Image sharpening section
200: cytogenetic layer detector
201: image normalization unit
202: cytoskeleton layer structure analysis section
203: Photoreceptor layer extracting unit
204:
300: Quantification of cytoskeleton layer
S100: Image input step
S150: image preprocessing step
S200: photocell layer detection step
S300: Quantification step of cytosol layer

Claims (24)

In an apparatus for detecting a photoreceptor layer in an ocular image,
An image input unit for receiving an ophthalmologic tomography image;
A photocell layer detector for detecting at least one photocell layer determined in advance in the ocular tomographic image; And
And a cytological cell layer quantifying unit for receiving the ocular disease information and correcting the erroneously detected region in the detected ocular cell layer according to the input ocular disease information to quantify the ocular cell layer. The method according to claim 1,
Wherein the ocularly tomographic image is an image photographed using an optical coherence tomography apparatus. The method according to claim 1,
Wherein the image input unit receives the 3-dimensional ocularly tomographic image or receives the ocular tomographic image including at least one 2-dimensional ocular tomographic image. The method according to claim 1,
Wherein the ocular tomographic image represents a layered structure of the photoreceptor layer including the macula of the retina, The method according to claim 1,
Further comprising an image preprocessing unit for performing at least one of performing normalization of a signal size of the ocularly tomographic image, noise elimination filtering, and image sharpening signal processing, and pre-processing the ocularly tomographic image ,
Wherein the cytosolic layer detecting unit detects the photocell layer in the pre-processed ocular tomographic image. [6] The apparatus of claim 5, wherein the image preprocessing unit comprises:
A normalization is performed to adjust a signal size of an image according to a predetermined standard,
Downsampling to reduce the image size,
Performing at least one of median filtering and Gaussian filtering to remove noise,
And performing image sharpening signal processing for enhancing the edge of the image. 5. The method of claim 4,
Wherein the cytosolic layer detector sets an inspection line in a direction perpendicular to the layered structure of the photocell layer in the ocular tomographic image and obtains an inspection line image signal component indicating a change in the image signal size on the inspection line, And the photocell layer is detected by analyzing the image signal component on the inspection line obtained. 8. The method of claim 7,
Wherein the cytosolic layer detecting unit analyzes the video signal component on the inspection line and laminates the detected cytoskeleton layer two-dimensionally or three-dimensionally, and detects the photocell layer in two or three dimensions in the ocular tomographic image Wherein the cytoskeletal layer detection device detects the cytoskeletal layer. 8. The method of claim 7,
Wherein the visual-field-layer detecting unit divides the image signal component on the inspection line into an inner segment and an outer segment, and based on the rising or falling degree of the image signal value of the image signal component on the inspection line, Wherein at least one of the second photocell layer and the third photocell layer is detected at the portion where the first photocell is detected at the end portion or at the portion where the second photocell is detected. 10. The method of claim 9,
Wherein the photodetector layer detection unit detects one local minimum value located between two local maximum values and the two local maximum values in the image signal component on the inspection line, Is divided into the above-mentioned knife-cutting part and the above-mentioned part of the outer circumference. 11. The imaging apparatus according to claim 10, wherein the photocell layer detector detects,
Wherein the first photocell layer is formed on the basis of a point where the degree of increase of the image signal value is maximum in the inner fold portion,
The second photocell layer is formed on the basis of a point where the degree of rise of the image signal value is maximum in the non-
And the third photocell layer is detected based on a point at which the degree of descent of the image signal value is maximum at the outer edge portion. 11. The method of claim 10,
Wherein the cytosolic layer detection unit calculates the rising or falling degree of the video signal value of the video signal component on the inspection line using the differential value with respect to the video signal component on the inspection line. The method according to any one of claims 9 to 12,
The first photocell layer may include an ILM (Inner Limiting Membrane) layer,
The second photocell layer may include an inner segment (IS) layer,
Wherein the third photocell layer is an RPE (Retinal Pigment Epithelium) layer. 9. The method of claim 8,
The eye disease information includes information on the types of eye diseases,
Wherein the cytologic layer quantification unit quantizes the at least one cytocentar layer selected according to the ocular disease information,
Wherein the photocell layer is discontinuous in the two-dimensional or three-dimensionally detected photocell layer in the photocell layer detecting section, or a region in which a change in the signal size is larger than a predetermined reference is detected as an erroneous detection region,
And corrects the erroneous detection region in the cytoskeleton layer to quantify the cytoskeleton layer. 15. The method of claim 14,
Wherein the cytological cell layer quantification unit corrects the erroneous detection region in the photocell layer,
Correcting the erroneous detection area using another photocell layer,
Or interpolates the erroneous detection region using the neighboring region of the erroneous detection region in the photoreceptor layer. 16. The method of claim 15, wherein the cytosolic layer quantification unit comprises:
Wherein the ILM layer and the IS layer of the photoreceptor layer detected by the photoreceptor layer detection unit correspond to a circular region that is a lesion region according to the macular hole when the type of ocular disease corresponds to macular hole (MH) Replacing the ILM layer with the IS layer,
When the type of ocular disease is one of diabetic macular edema (DME) or branch retinal vein occlusion (BRVO) or age-related macular degeneration (AMD) A discontinuity section having a change in the video signal size that is greater than or equal to a predetermined reference is detected for the IS layer or the RPE layer detected by the cytosolic layer detection section and the discontinuity section is interpolated using the interpolation method when the discontinuity section is detected Wherein the cell layer is a cell. 8. The method of claim 7,
The eye disease information includes information on the types of eye diseases,
Wherein the cytologic layer quantification unit determines whether or not the position of the cytosolic layer is erroneously detected with respect to at least one cytoskdial layer selected according to the eye disease information,
And corrects the position of the photoreceptor layer when the photoreceptor is erroneously detected as a result of the determination, thereby quantifying the photoreceptor layer. 18. The method of claim 17, wherein the cytosolic layer quantification unit comprises:
If the type of ocular disease corresponds to high myopia (HM), it is determined whether at least one of the ILM layer, the IS layer, and the RPE layer of the photoreceptor layer has a section deviating from the retinal tomography image Correcting the location of the photoreceptor layer in the section of the photoreceptor layer outside the retinal tomography image according to the determination result,
Wherein when the type of ocular disease corresponds to an epi-retinal membrane (ERM), a degree of rise of an image signal value at an end portion of the image signal component of the inspection line is calculated, , And corrects the candidate layer closer to the retinal macular region in the two detected candidate layers to the ILM layer. ≪ RTI ID = 0.0 > In an apparatus for detecting a photoreceptor layer in an ocular image,
An image input unit receiving an ocular tomography image showing a layered structure of a photoreceptor layer including a macula of the retina;
The inspection line is set in a direction perpendicular to the layered structure of the photoreceptor layer in the ocular tomography image, an inspection line image signal component indicating a change in the image signal size on the inspection line is obtained, And a photocell layer detector for analyzing a phase image signal component and detecting at least one of the predetermined photocell layers determined in advance. 20. The method of claim 19,
Further comprising an image preprocessing unit for normalizing the signal size of the ocularly tomographic image, performing noise elimination filtering, and performing image sharpening signal processing to pre-process the ocular tomographic image,
Wherein the cytosolic layer detecting unit detects the photocell layer in the pre-processed ocular tomographic image. 20. The method of claim 19,
Wherein the visual-field-layer detecting unit divides the image signal component on the inspection line into an inner segment and an outer segment, and based on the rising or falling degree of the image signal value of the image signal component on the inspection line, Wherein at least one of the second photocell layer and the third photocell layer is detected at the portion where the first photocell is detected at the end portion or at the portion where the second photocell is detected. 22. The method of claim 21,
Wherein the photodetector layer detection unit detects one local minimum value located between two local maximum values and the two local maximum values in the image signal component on the inspection line, Is divided into the above-mentioned knife-cutting part and the above-mentioned part of the outer circumference. 22. The method of claim 21,
The first photocell layer may include an ILM (Inner Limiting Membrane) layer,
The second photocell layer may include an inner segment (IS) layer,
The third photocell layer is an RPE (Retinal Pigment Epithelium) layer,
The visual-field-layer detecting unit may detect the visual-
Wherein the first photocell layer is formed on the basis of a point where the degree of increase of the image signal value is maximum in the inner fold portion,
The second photocell layer is formed on the basis of a point where the degree of rise of the image signal value is maximum in the non-
And the third photocell layer is detected based on a point at which the degree of descent of the image signal value is maximum at the outer edge portion. A method for detecting a photoreceptor layer in an ocular image,
An image input step of receiving an ophthalmologic tomography image;
A photocell layer detecting step of detecting at least one predetermined photocell layer determined in the ocular tomographic image; And
And a cytocentric layer quantifying step of inputting the ocular disease information and correcting the erroneously detected region in the detected ocular cell layer according to the input ocular disease information to quantify the ocular cell layer.

Images