CN109118499B - Layer segmentation method of coherent light tomography image based on backtracking shortest path algorithm - Google Patents
Layer segmentation method of coherent light tomography image based on backtracking shortest path algorithm Download PDFInfo
- Publication number
- CN109118499B CN109118499B CN201810756901.1A CN201810756901A CN109118499B CN 109118499 B CN109118499 B CN 109118499B CN 201810756901 A CN201810756901 A CN 201810756901A CN 109118499 B CN109118499 B CN 109118499B
- Authority
- CN
- China
- Prior art keywords
- image
- node
- weight
- backtracking
- shortest path
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 59
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 36
- 230000011218 segmentation Effects 0.000 title claims abstract description 33
- 230000001427 coherent effect Effects 0.000 title claims abstract description 17
- 238000003325 tomography Methods 0.000 title claims abstract description 17
- 238000004364 calculation method Methods 0.000 claims description 9
- 230000004044 response Effects 0.000 claims description 9
- 230000001186 cumulative effect Effects 0.000 claims description 4
- 230000009466 transformation Effects 0.000 claims description 4
- 238000012014 optical coherence tomography Methods 0.000 description 19
- 210000001525 retina Anatomy 0.000 description 9
- 230000036285 pathological change Effects 0.000 description 4
- 230000002207 retinal effect Effects 0.000 description 4
- 238000006243 chemical reaction Methods 0.000 description 3
- 230000003902 lesion Effects 0.000 description 3
- 231100000915 pathological change Toxicity 0.000 description 3
- 238000012567 pattern recognition method Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 208000002367 Retinal Perforations Diseases 0.000 description 2
- 238000003491 array Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 208000029233 macular holes Diseases 0.000 description 2
- 230000003595 spectral effect Effects 0.000 description 2
- 208000001344 Macular Edema Diseases 0.000 description 1
- 206010025415 Macular oedema Diseases 0.000 description 1
- 206010064930 age-related macular degeneration Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000012472 biological sample Substances 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000011423 initialization method Methods 0.000 description 1
- 238000005305 interferometry Methods 0.000 description 1
- 208000002780 macular degeneration Diseases 0.000 description 1
- 201000010230 macular retinal edema Diseases 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000004660 morphological change Effects 0.000 description 1
- 208000020911 optic nerve disease Diseases 0.000 description 1
- 230000007170 pathology Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000004256 retinal image Effects 0.000 description 1
- 239000000523 sample Substances 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000004304 visual acuity Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/12—Edge-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10101—Optical tomography; Optical coherence tomography [OCT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30041—Eye; Retina; Ophthalmic
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Eye Examination Apparatus (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a layer segmentation method of a coherent light tomography image based on a backtracking shortest path algorithm, which comprises the following steps: s1, denoising the original image; s2, calculating the weights of all nodes in the graph to form a weight graph for the image after the noise is removed; s3, defining a starting point and an end point of the shortest path algorithm on the weight graph by adopting a method of automatically initializing the end points; s4, calculating and finding the minimum accumulated cost path between the starting point and the end point by using a shortest path algorithm with backtracking on the weight graph; s5, removing the auxiliary columns added when the end points are defined in the obtained minimum accumulated cost path to obtain a layer segmentation result of one layer in the coherent light tomography image; limiting the image search area, and repeating the steps S3-S5 to finally obtain the segmentation results of all layers in the coherent light tomography image. The segmentation method avoids the short circuit problem in the traditional method, is suitable for segmenting normal and lesion-affected coherent tomography images, and has high accuracy and flexibility.
Description
Technical Field
The invention relates to the technical field of coherent light tomographic image processing, in particular to a layer segmentation method of a coherent light tomographic image based on a backtracking shortest path algorithm.
Background
Oct (optical Coherence tomography), also known as optical Coherence tomography, is a new tomographic technique with high resolution and tomographic features on biological internal structures. OCT is able to obtain sub-surface images at near microscopic resolution because it uses light waves as a detection means rather than acoustic waves. The wavelength of light wave is shorter than that of sound wave, and the propagation speed is faster than that of sound wave. OCT was developed based on low coherence interference techniques and has made tremendous progress in recent decades. One generation of OCT, also known as time-domain OCT (TD-OCT), acquires 400 a scans per second with an axial resolution of 8 to 10 microns. Spectral domain OCT (SD-OCT) appeared in 2011. The imaging speed of the spectral domain OCT equipment which is commercially used at present is 10 to 100 times faster than that of the time domain OCT equipment, and the axial resolution range is 3 to 7 micrometers. Recently, swept source OCT (SS-OCT) has been developed, which can acquire over a hundred thousand a-scans per second, with axial resolution up to 3 microns.
OCT can image biological tissue in real time without the need to prepare a sample. It does not involve ion radiation and has no harm to human health. OCT has rapidly developed in many fields and can provide high-resolution cross-sectional images from backscatter profiles of biological samples by using low coherence interferometry. This allows pathological and morphological changes in the retina to be observed during the progression of the disease or before the patient recognizes visual changes (visual acuity). Over the past two decades, OCT has become a well established imaging method widely used by ophthalmologists in the diagnosis of retinal and optic nerve diseases. Distortion of the retinal layer is an important signal for experts in the diagnosis of age-related macular degeneration, macular edema, and macular hole. However, manual layer segmentation is typically a time consuming and poorly repeatable process. Therefore, an automatic or semi-automatic segmentation method of the OCT image is necessary.
Some automatic or semi-automatic OCT slice segmentation methods are currently available, which can be roughly classified into three types, a scan-based method, a scan-B based method, and a three-dimensional volume data-based method. The a-scan based method can detect the peak or valley of intensity on the boundary on each a-scan section, and then connect the detected points by using model fitting technique to form a smooth and continuous boundary. However, the a-scan based method mostly does not consider information around the detection point and is easily affected by noise, and thus, the accuracy and robustness of the a-scan based method are limited. Common methods based on B-scan include active contour methods, graph search methods based on shortest paths, and statistical model methods. B-scan methods are generally superior to a-scan methods, however, they still may fail to detect diseased retinal structures. The existing segmentation method based on volume data mainly adopts a 3D map-based method and a pattern recognition method, but the methods are generally high in calculation cost, and the pattern recognition method generally needs training data which is manually segmented by experts to learn a feasible classification model. These pattern recognition methods are also affected in terms of accuracy and efficiency. Although image algorithms applicable to normal images or limited pathologies have made great progress, segmentation of retinal image layers with significant pathological changes remains a challenge. For example, in the macular hole, when the central region is severely deformed, the above method cannot accurately perform segmentation.
The algorithm for solving the shortest path problem in the graph is called as a shortest path algorithm, and the most commonly used shortest path algorithms are Dijkstra algorithm, Bellman-Ford algorithm and Floyd-Warshall algorithm. In the shortest path method, it extracts a curve-like structure in an image with the smallest cumulative cost through a communication path between two given points (a start point and an end point). However, the traditional shortest path algorithm has a short circuit problem, the improved shortest path algorithm with backtracking can effectively avoid the short path problem, and the method is suitable for segmentation images of the pathological change retina layer structure.
Disclosure of Invention
In order to overcome the defects of the technology, the invention provides a layer segmentation method of a coherent light tomography image based on a backtracking shortest path algorithm.
The technical scheme adopted by the invention for overcoming the technical problems is as follows:
a layer segmentation method of a coherent light tomography image based on a backtracking shortest path algorithm comprises the following steps:
s1, denoising the original image;
s2, calculating the weights of all nodes in the graph to form a weight graph for the image after the noise is removed;
s3, defining a starting point and an end point of the shortest path algorithm on the weight graph by adopting a method of automatically initializing the end points;
s4, calculating and finding the minimum accumulated cost path between the starting point and the end point by using a shortest path algorithm with backtracking on the weight graph;
s5, removing the auxiliary columns added when the end points are defined in the obtained minimum accumulated cost path to obtain a layer segmentation result of one layer in the coherent light tomography image;
limiting the image search area, and repeating the steps S3-S5 to finally obtain the segmentation results of all layers in the coherent light tomography image.
Preferably, in the step S1, the BM3D algorithm is adopted for denoising, and the method is fast and convenient.
Preferably, in step S2, the node weight calculation includes a gray scale weight, a gradient weight, and a maximum response value weight after Gabor transformation of the image.
Preferably, in step S4, the following formula is adopted to calculate and find the minimum cumulative cost path between the starting point and the ending point on the weight map by using the shortest path algorithm with backtracking:
where N represents the current image node,representing the accumulated cost of the backtracking length l of the current image node to the previous point N-l +1 towards the starting node; (ii) a
Node WiThe cost calculation formula is as follows:
Wi=r*Wi gray+s*Wi grad+t*Wi gabor*Dif
wherein, Wi grayIs the image gray value of the node, Wi gradIs a gradient weight, Wi gaborThe maximum Gabor response value of a node, r, s, t, is three constantsThe number parameter, r + s + t, is 1.
The invention has the beneficial effects that:
the method of the invention firstly removes noise from an OCT image of a lesion, calculates all node weight maps in the map based on the noise-reduced image, then defines a starting point and an end point by an automatic endpoint initialization method, and finally calculates the minimum accumulated cost path between the starting point and the end point on the weight map by using a shortest path algorithm with backtracking as a retina layer segmentation result. The segmentation method of the invention has the following advantages:
(1) high accuracy. The method uses a shortest path algorithm with backtracking, is used in the technical field for the first time, carries out layer structure segmentation on the lesion coherence tomography image, avoids the problem of short circuit in the traditional method, and obtains better segmentation effect and performance compared with the existing method.
(2) Flexibility. The method is also suitable for the layer structure segmentation of the normal coherence tomography image, can obtain a better segmentation result image on the normal image, and has higher flexibility and adaptability.
Drawings
FIG. 1 is a schematic flow chart of an embodiment of the present invention.
Fig. 2 is a schematic view of the retinal layer structure.
Fig. 3(a) is the original before Gabor conversion, fig. 3(b), fig. 3(c), and fig. 3(d) are response graphs in the 30-degree direction, the horizontal direction, and the vertical direction, respectively, fig. 3(e) is the maximum response graph after Gabor conversion, and fig. 3(f) is the direction information graph after Gabor conversion.
Fig. 4 is an original before division.
Fig. 5 is a graph of the layer division result of fig. 4.
Fig. 6 is a schematic diagram of automatically initializing the start and end points.
Detailed Description
In order to facilitate a better understanding of the invention for those skilled in the art, the invention will be described in further detail with reference to the accompanying drawings and specific examples, which are given by way of illustration only and do not limit the scope of the invention.
The invention discloses a layer segmentation method of a coherent light tomography image based on a backtracking shortest path algorithm, and the specific implementation process of the method is described below by taking a pathological change retina layer as an example and combining with accompanying drawings 1-6.
Step one, denoising an original image
The retina image data used in the invention has noise, which affects the final image segmentation result, and many image denoising methods exist, and the embodiment adopts BM3D algorithm to perform fast and convenient denoising processing to obtain the denoised image.
The BM3D algorithm: the method comprises the steps of firstly dividing an image into blocks with a certain size, combining two-dimensional image blocks with similar structures together to form three-dimensional arrays according to the similarity between the image blocks, then processing the three-dimensional arrays by a combined filtering method, and finally returning the processed result to the original image through inverse transformation, thereby obtaining the denoised image.
Step two, calculating the node weight to obtain a weight graph
After the denoised image is obtained in the first step, the second step needs to calculate all node weights on the denoised image, in this embodiment, the node weights are calculated by using image intensity, gradient information and Gabor direction information, which is specifically as follows:
first, an intensity weight of the image is calculated. The image intensity represents the intensity of a single-channel image pixel, in our dataset, the image is a gray scale image, so the intensity is the gray scale of the pixel, in the retina OCT image, the calculation formula of the intensity weight of this embodiment is as follows:
where I (I, j) represents the image gray scale value of the node in the original image, I, j is the position of the pixel in the image, and k is a constant value set to 1.
Second, the weight of the gradient information is calculated. The gradient image is obtained by convolving the original image with a filter, the image gradient can be used to extract the edge information of the image, in this embodiment, the layer in the retina OCT image can be used as the edge information, so the gradient information is used, and the calculation formula of the image gradient weight is as follows:
dx(i,j)=I(i+1,j)-I(i,j) (3)
where I is the image, I, j are the coordinates of the pixels, and there are two types of edge considerations: dark to light edges and light to dark edges, we calculate dy (i, j) as shown above.
And finally, calculating the weight of the Gabor direction information. The Gabor transform is a special case of a short-time fourier transform, which is commonly used to determine the sinusoidal frequency and phase content of a signal as it varies over time, and can be used to extract features in different directions. In the method, a Gabor transform is applied to each pixel to calculate its response value in different directions, and then the calculation is performed using a Gabor maximum response value as a weight.
When calculating the weight of a node, the three weight calculation methods described above are combined as follows:
Wi=r*Wi gray+s*Wi grad+t*Wi gabor*Dif (5)
wherein, Wi grayImage grey scale weight of node, Wi gradIs a gradient weight, Wi gaborFor the maximum response value of a node, r, s, t are three constant parameters, and the sum of r + s + t is 1.
Step three, automatically initializing algorithm propagation starting point and end point
FIG. 6 shows an example image with an auto-initialization technique. Firstly, adding a row of weight values to the left side and the right side of the weight graph obtained in the step two, wherein the weight values of the two rows are set as the minimum value of the whole weight graph; then, the starting point and the end point are automatically initialized to the upper left corner node and the lower right corner node, as shown in fig. 6, and the leftmost column and the rightmost column are the added auxiliary weight minimum value columns.
Step four, calculating the minimum accumulated cost path by using the shortest path algorithm with backtracking on the weight graph
Because of the short circuit problem, the distorted retina layer structure cannot be accurately segmented by the original shortest path algorithm, and therefore, the problem is solved by adopting a backtracking method in the embodiment. When the accumulated cost is calculated between the starting point and the end point of the weight graph, backtracking a certain length from the current node to the starting point direction, and replacing the node cost with the backtracked short edge accumulated cost. In doing so, the problem of short circuits can be avoided, as follows:
where N is the current image node,and representing the accumulated cost of the backtracking length l of the current image node to the previous point.
In addition, the implementation method also defines a new weight component by utilizing Gabor direction information to limit the propagation direction of the algorithm. In the layer segmentation task of OCT images, the direction of the layers is usually horizontal, as shown in fig. 2, and this a priori information can be exploited to improve the robustness of the segmentation method. When the gabor transform is applied to an image, each node has a principal direction value, which is used in the method to limit the propagation direction, as shown in fig. 3 (f). In the algorithm propagation process, the average direction value of the traced short side is calculated, the average direction is used as a propagation trend, the direction value of the current node is compared with the average direction value, and the propagation of the next node is determined.
Step five, removing the auxiliary column added when the end point is defined in the obtained minimum accumulated cost path to obtain the layer segmentation result of one layer in the coherent light tomography image
As shown in fig. 6, on the weight graph after the auxiliary columns are added, lines represent an example result after the method of the present embodiment is used, and then the left and right columns of auxiliary columns that are added are deleted, so as to obtain a segmentation result of the ILM layer in the lesion retina layer structure.
After the ILM layer is segmented, the image search area is limited, and the above steps three to five are repeated to obtain the segmentation results of the remaining layers as shown in fig. 2.
In the description of the present invention, the terms "horizontal", "vertical", "upper", "lower", "left", "right", and the like indicate orientations or positional relationships based on those shown in the drawings, and are only for the purpose of describing the present invention and simplifying the description, and do not indicate or imply that the structures referred to must have a specific orientation, be constructed in a specific orientation, and be operated, and thus, should not be construed as limiting the present invention.
The foregoing merely illustrates the principles and preferred embodiments of the invention and many variations and modifications may be made by those skilled in the art in light of the foregoing description, which are within the scope of the invention.
Claims (3)
1. A layer segmentation method of a coherent light tomography image based on a shortest path backtracking algorithm is characterized by comprising the following steps:
s1, denoising the original image;
s2, calculating the weights of all nodes in the graph to form a weight graph for the image after the noise is removed;
s3, defining a starting point and an end point of the shortest path algorithm on the weight graph by adopting a method of automatically initializing the end points;
s4, calculating and finding the minimum accumulated cost path between the starting point and the end point by using a shortest path algorithm with backtracking on the weight graph; calculating and finding the minimum accumulated cost path between the starting point and the end point by using a shortest path algorithm with backtracking on the weight graph by adopting the following formula:
where N represents the current image node,representing the cumulative sum of node weight values from the start node start to the current image node N on the path,representing the cumulative sum of node weight values from the start node start to the image node N-l on the path,representing the accumulated cost of the backtracking length l of the current image node to the previous point N-l +1 towards the starting node;
cost W of node iiThe calculation formula of (a) is as follows:
Wi=r*Wi gray+s*Wi grad+t*Wi gabor*Dif
wherein, Wi grayIs the image gray value of the node, Wi gradIs a gradient weight, Wi gaborThe maximum response value weight after Gabor transformation of the node is shown, r, s and t are three constant parameters, r + s + t is 1, and Dif represents the direction deviation of the path;
s5, removing the auxiliary columns added when the end points are defined in the obtained minimum accumulated cost path to obtain a layer segmentation result of one layer in the coherent light tomography image;
limiting the image search area, and repeating the steps S3-S5 to finally obtain the segmentation results of all layers in the coherent light tomography image.
2. The method according to claim 1, wherein in step S1, the BM3D algorithm is used for denoising.
3. The method according to claim 1, wherein in step S2, the node weight calculation includes a gray scale weight, a gradient weight, and a maximum response value weight after Gabor transformation of the image.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810756901.1A CN109118499B (en) | 2018-07-11 | 2018-07-11 | Layer segmentation method of coherent light tomography image based on backtracking shortest path algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810756901.1A CN109118499B (en) | 2018-07-11 | 2018-07-11 | Layer segmentation method of coherent light tomography image based on backtracking shortest path algorithm |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109118499A CN109118499A (en) | 2019-01-01 |
CN109118499B true CN109118499B (en) | 2020-06-16 |
Family
ID=64862591
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810756901.1A Active CN109118499B (en) | 2018-07-11 | 2018-07-11 | Layer segmentation method of coherent light tomography image based on backtracking shortest path algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109118499B (en) |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104143190A (en) * | 2014-07-24 | 2014-11-12 | 东软集团股份有限公司 | Method and system for partitioning construction in CT image |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107332770B (en) * | 2017-08-18 | 2020-05-19 | 苏州浪潮智能科技有限公司 | Method for selecting routing path of necessary routing point |
-
2018
- 2018-07-11 CN CN201810756901.1A patent/CN109118499B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104143190A (en) * | 2014-07-24 | 2014-11-12 | 东软集团股份有限公司 | Method and system for partitioning construction in CT image |
Non-Patent Citations (2)
Title |
---|
Automatic segmentation of seven retinal layers in SDOCT images congruent with expert manual segmentation;Stephanie J. Chiu, et al.;《OPTICS EXPRESS》;20100830;第18卷(第18期);第19413-19428页 * |
基于超高分辨率OCT图像的视网膜层状结构研究;徐肃仲等;《国际眼科杂志》;20140831;第14卷(第8期);第1494-1497页 * |
Also Published As
Publication number | Publication date |
---|---|
CN109118499A (en) | 2019-01-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
KR101017611B1 (en) | System and method for extracting anatomical feature | |
US9418423B2 (en) | Motion correction and normalization of features in optical coherence tomography | |
Kafieh et al. | A review of algorithms for segmentation of optical coherence tomography from retina | |
US11887319B2 (en) | Systems and methods for reducing artifacts in oct angiography images | |
Novosel et al. | Loosely coupled level sets for simultaneous 3D retinal layer segmentation in optical coherence tomography | |
CN106558030B (en) | Choroid segmentation method in three-dimensional large-visual-field swept-frequency optical coherence tomography | |
Golabbakhsh et al. | Vessel‐based registration of fundus and optical coherence tomography projection images of retina using a quadratic registration model | |
US9514513B2 (en) | Establishing compatibility between two- and three-dimensional optical coherence tomography scans | |
US20150371400A1 (en) | Segmentation and identification of closed-contour features in images using graph theory and quasi-polar transform | |
US20210012494A1 (en) | Methods for detection and enhanced visualization of pathologies in a human eye | |
US20210272291A1 (en) | Method and computer program for segmentation of optical coherence tomography images of the retina | |
Lang et al. | Segmentation of retinal OCT images using a random forest classifier | |
Kolar et al. | Registration of 3D retinal optical coherence tomography data and 2D fundus images | |
Wu et al. | Automated fovea detection in spectral domain optical coherence tomography scans of exudative macular disease | |
Rabbani et al. | Obtaining thickness maps of corneal layers using the optimal algorithm for intracorneal layer segmentation | |
Eladawi et al. | Optical coherence tomography: A review | |
Gan et al. | Robust layer segmentation of esophageal OCT images based on graph search using edge-enhanced weights | |
Shi et al. | Automated choroid segmentation in three-dimensional 1-μ m wide-view OCT images with gradient and regional costs | |
Pan et al. | Segmentation guided registration for 3d spectral-domain optical coherence tomography images | |
CN109118499B (en) | Layer segmentation method of coherent light tomography image based on backtracking shortest path algorithm | |
US10441164B1 (en) | Correction of decorrelation tail artifacts in a whole OCT-A volume | |
Zhang et al. | Fast retinal layer segmentation of spectral domain optical coherence tomography images | |
Stankiewicz et al. | Matching 3d oct retina images into super-resolution dataset | |
US11481905B2 (en) | Atlas for automatic segmentation of retina layers from OCT images | |
Sharif et al. | Extraction and analysis of RPE layer from OCT images for detection of age related macular degeneration |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |