CN115690124A - High-precision single-frame fundus fluorography image leakage area segmentation method and system - Google Patents
High-precision single-frame fundus fluorography image leakage area segmentation method and system Download PDFInfo
- Publication number
- CN115690124A CN115690124A CN202211363937.6A CN202211363937A CN115690124A CN 115690124 A CN115690124 A CN 115690124A CN 202211363937 A CN202211363937 A CN 202211363937A CN 115690124 A CN115690124 A CN 115690124A
- Authority
- CN
- China
- Prior art keywords
- image
- leakage
- fundus
- segmentation
- region
- 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.)
- Granted
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Eye Examination Apparatus (AREA)
Abstract
The invention discloses a method and a system for segmenting a leakage area of a high-precision single-frame fundus fluorescence angiography image based on a Gaussian mixture model, wherein the method comprises the following steps: s1, screening a report of macular leakage or optic disc leakage; s2, selecting a later period t of radiography 1 To t 2 The minute fundus fluorography image is taken as a working data set; s3, intercepting an image containing a macular leakage or optic disc leakage area; s4, generating a retinal blood vessel mask; s5, removing the image I 0 A vascular portion of superior; s6, carrying out desalination treatment; s7, segmenting the image X' by adopting a method based on a Gaussian mixture model; and S8, optimizing the preliminary leakage segmentation image. The method and the system for segmenting the leakage area of the high-precision single-frame fundus fluorography image based on the Gaussian mixture model can realize high-precision segmentation of the single-frame fundus fluorography image and have potential for auxiliary diagnosis of fundus diseasesThe medical value of (1).
Description
Technical Field
The invention relates to the field of medical image processing, in particular to a high-precision single-frame fundus fluorography image leakage area segmentation method and system.
Background
Fundus fluorography is the "gold standard" for diagnosing the impaired state of the early retinal barrier, and one of its important functions is to image the fundus leak. To perform further analysis of fundus leakage on the image, the fundus leakage region needs to be segmented. At present, the division of the fundus oculi leakage area is mostly realized in a manual mode in clinic. The manual segmentation mode is time-consuming and labor-consuming, and has strong subjectivity. Therefore, it is particularly important to develop an automatic fundus leakage image segmentation method.
Existing automated fundus leak segmentation methods are implemented primarily through analysis of multiple frames of fundus fluorography images, which require accurate registration of the multiple frames of fundus fluorography images. However, such accurate registration is complicated to achieve by algorithms and has a certain probability of registration failure. Therefore, the single-frame fundus fluorescence leakage region segmentation method has great advantages. However, the existing single-frame fundus fluorescence leakage region division method has the problems of low division accuracy or long training.
Disclosure of Invention
The invention aims to solve the technical problem of providing a method and a system for segmenting a leakage area of a high-precision single-frame fundus fluorography image based on a Gaussian mixture model, aiming at the defects in the prior art.
In order to solve the technical problems, the technical scheme adopted by the invention is as follows: a high-precision single-frame fundus fluorescence angiography image leakage area segmentation method comprises the following steps:
s1, screening out a report of macular leakage or optic disc leakage from collected fundus fluorography image reports;
s2, selecting a later contrast period t from the reports screened in the step S1 1 To t 2 The minute fundus fluorography image is taken as a working data set;
s3, preprocessing the fundus fluorescence angiography image in the working data set: intercepting an image containing a macular leakage or optic disc leakage area to obtain a preprocessed fundus fluorography image I 0 ;
S4, utilizing the image I obtained in the step S3 0 Generating a retinal vascular mask;
s5, removing the image I by using the retinal blood vessel mask obtained in the step S4 0 Obtaining a blood vessel-free image W, and carrying out Gaussian blur processing on the blood vessel-free image W to obtain an image W';
s6, carrying out desalination treatment on the background of the image W': removing the background of the image W 'to obtain a non-blood vessel image X, and then endowing the non-blood vessel image X with spatial Gaussian distribution weight to obtain an image X' after background desalination;
s7, segmenting the image X' by adopting a method based on a Gaussian mixture model to obtain a primary leakage segmentation image;
and S8, optimizing the preliminary leakage segmentation image obtained in the step S7 to obtain a final leakage segmentation image.
Preferably, the step S4 specifically includes:
s4-1, utilizing Sobel operator to carry out comparison on image I 0 Processing to obtain image I 0 Is represented as:
wherein x, y represent pixels;
s4-2, processing the gradient image S (x, y) by respectively adopting corrosion operation and expansion operation to obtain a corroded imageErode (S (x, y)) and the expanded image Dilate (S (x, y)), and then the region where the pixel difference between the image Dilate (S (x, y)) and the image Erode (S (x, y)) is larger than P1 is regarded as the region S where the blood vessel exists RBVR The calculation formula is as follows:
S RBVR ={(x,y)|Dilate(S(x,y))-Erode(S(x,y))>P1};
s4-3, to region S RBVR Performing a closing operation to fill the isolated region to obtain a retinal blood vessel mask S FBVR Expressed as:
S FBVR =Close(S RBVR )。
preferably, the step S5 specifically includes:
s5-1, and comparing the fundus fluorescent contrast image I obtained in the step S3 0 And performing opening operation processing to remove blood vessels U on the fundus fluorography image and obtain a blood vessel-free image with an affected leakage area, wherein U is expressed as:
U=Open(I);
s5-2, in the area covered by the retinal blood vessel mask, assigning the affected non-vascular image of the leakage area to the fundus fluorography image I 0 A non-vascular image W in which the leakage area is not affected is obtained, and is expressed as:
s5-3, performing Gaussian blur processing on the non-vascular image W without the influence of the leakage area to fade the mask edge artifact and obtain a processed image W', which is expressed as:
W'=Gaussian Blur(W)。
preferably, the step S6 specifically includes:
s6-1, carrying out corrosion operation on the non-blood vessel image W 'without the affected leakage area obtained in the step S5 to obtain a background image Erode (W') with the leakage area eliminated, and then subtracting the background image Erode (W ') from the image W' to obtain a non-blood vessel image X without the background, wherein the non-blood vessel image X is expressed as:
X=W'-Erode(W');
s6-2, endowing the background-removed non-vascular image with spatial Gaussian distribution weight to obtain an image X' after background fading treatment, and expressing as follows:
X'(x,y)=X(x,y)×Gaussian Distribution(x,y)。
preferably, the step S7 specifically includes:
s7-1, down-sampling the image X' obtained in the step S6 to obtain a down-sampled image D (X, y) which is expressed as:
D(x,y)=Downsample(X'(x,y));
s7-2, inputting the down-sampled image D (x, y) into a Gaussian mixture model and calculating the mean value M of the mu values of the image Gaussian mixture model:
M=Mean(Gaussian Mixture Model μ (D(x,y)));
s7-3, taking alpha times of the mean value M as a threshold, and carrying out threshold segmentation on the image X' obtained in the step S6 to obtain a primary leakage segmentation image S RLR Expressed as:
S RLR ={(x,y)|X'(x,y)>α×M};
wherein alpha is more than 0 and less than 1.
Preferably, the step S8 specifically includes:
first, the preliminary leakage segmentation image S obtained in step S7 is calculated RLR The area of each connected region; then, the maximum value S of the area of the communication region is calculated max (ii) a Finally, the area is less than beta S max Segmentation of the image S from the preliminary leakage RLR And (5) removing to obtain a final leakage segmentation image, wherein beta is more than 0 and less than 1.
Preferably, wherein t is 1 =5,t 2 =6,P1=50,α=0.75,β=0.02。
The invention also provides a high-precision single-frame fundus fluorography image leakage area segmentation system which adopts the method to segment the fundus fluorography image leakage area.
The invention also provides a storage medium having a computer program stored thereon, characterized in that the program is adapted to carry out the method as described above when executed.
The invention also provides a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the method as described above when executing the computer program.
The invention has the beneficial effects that: the method and the system for segmenting the leakage area of the high-precision single-frame fundus fluorography image based on the Gaussian mixture model can realize high-precision segmentation of the single-frame fundus fluorography image, and have potential medical value for auxiliary diagnosis of fundus diseases.
Drawings
FIG. 1 is a flow chart of a high-precision single-frame fundus fluorography image leakage region segmentation method of the present invention;
FIG. 2 is two types of fundus fluorescence angiography images applied in example 1 of the present invention;
FIG. 3 shows the results of the segmentation of the leakage in the clinical data set by the method of example 1 of the present invention and the comparison with the results of the manual segmentation by experts.
Detailed Description
The present invention is further described in detail below with reference to examples so that those skilled in the art can practice the invention with reference to the description.
It will be understood that terms such as "having," "including," and "comprising," as used herein, do not preclude the presence or addition of one or more other elements or groups thereof.
Example 1
Referring to fig. 1, the present embodiment provides a high-precision single-frame fundus fluorography image leakage region segmentation method, which includes the following steps:
s1, screening out a report of macular leakage or optic disc leakage from collected fundus fluorography image reports;
in this example, the fluoroscopic images acquired were from 10 affected eyes at the third national hospital, in Changzhou, 5 of which had macular leaks and 5 of which had optic disc leaks. Fundus fluorescein angiography was performed by a Heidelberg confocal fundus angiography (Spectralis HRA).
S2, selecting fundus fluorescence angiography images in the later period of angiography of 5-6 minutes from the reports screened in the step S1 as a working data set;
s3, preprocessing the fundus fluorescence angiography image in the working data set: intercepting an image containing a macular or optic disc leakage area to obtain a preprocessed fundus fluorography image I 0 ;
S4, utilizing the image I obtained in the step S3 0 Generating a retinal vascular mask; the method specifically comprises the following steps:
s4-1, utilizing Sobel operator to pair image I 0 Processing to obtain image I 0 Is represented as:
wherein x, y represent pixels;
s4-2, processing the gradient image S (x, y) by respectively adopting an erosion operation and an expansion operation to obtain an eroded image Erode (S (x, y)) and an expanded image Diate (S (x, y)), and then determining the area where the pixel difference value between the image Diate (S (x, y)) and the image Erode (S (x, y) is more than 50 as an area S where the blood vessel exists RBVR (Rough Blood Vessels Region) with the formula:
S RBVR ={(x,y)|Dilate(S(x,y))-Erode(S(x,y))>50};
wherein, the die represents the swell operation, the anode represents the corrosion operation;
s4-3, for region S RBVR A closing operation process is performed to fill the isolated region (in the present embodiment, a region having a radius of less than 5 pixels is used as the isolated region), resulting in a retinal blood vessel mask S FBVR (Fine Blood Vessels Region) expressed as:
S FBVR =Close(S RBVR )。
s5, removing the image I by using the retinal blood vessel mask obtained in the step S4 0 Obtaining a non-vascular image W, and proceeding to the non-vascular image WPerforming Gaussian blur processing to obtain an image W';
the method specifically comprises the following steps:
s5-1, and for the fundus fluorography image I obtained in the step S3 0 And performing opening operation processing to remove blood vessels U on the fundus fluorography image and obtain a blood vessel-free image with an affected leakage area, wherein U is expressed as:
U=Open(I);
this operation, although removing the blood vessel portion on the image, affects the leak region on the image; thus, a vessel-free image is obtained in which the leak region is affected;
s5-2, in the area covered by the retinal blood vessel mask, assigning the affected non-vascular image of the leakage area to the fundus fluorography image I 0 A non-vascular image W in which the leakage area is not affected is obtained, and is expressed as:
s5-3, performing Gaussian blur processing on the non-vascular image W without the influence of the leakage area to fade the mask edge artifact and obtain a processed image W', which is expressed as:
W'=Gaussian Blur(W)。
s6, performing desalination treatment on the background of the image W': removing the background of the image W 'to obtain a non-blood vessel image X, and then endowing the non-blood vessel image X with spatial Gaussian distribution weight to obtain an image X' after background desalination;
the method specifically comprises the following steps:
s6-1, carrying out corrosion operation on the non-blood vessel image W 'without the affected leakage area obtained in the step S5 to obtain a background image Erode (W') with the eliminated leakage area, and then subtracting the background image Erode (W ') from the image W' to obtain a non-blood vessel image X without the background, wherein the non-blood vessel image X without the background is represented as:
X=W'-Erode(W');
s6-2, endowing the background-removed non-vascular image with spatial Gaussian distribution weight to obtain an image X 'after background desalination for further processing, and the image X' is expressed as:
X'(x,y)=X(x,y)×Gaussian Distribution(x,y)。
s7, segmenting the image X' by adopting a method based on a Gaussian mixture model to obtain a primary leakage segmentation image;
the method specifically comprises the following steps:
s7-1, down-sampling the image X' obtained in the step S6 to obtain a down-sampled image D (X, y) which is expressed as:
D(x,y)=Downsample(X'(x,y));
s7-2, inputting the down-sampled image D (x, y) into a Gaussian mixture model and calculating the mean value M of the mu values of the image Gaussian mixture model:
M=Mean(Gaussian Mixture Model μ (D(x,y)));
s7-3, performing threshold segmentation on the image X' obtained in the step S6 by taking 0.75 times of the mean value M as a threshold to obtain a primary leakage segmentation image S RLR Expressed as:
S RLR ={(x,y)|X'(x,y)>0.75×M};
s8, optimizing a segmentation result: optimizing the preliminary leakage segmentation image in the step S7 to obtain a final leakage segmentation image;
the method specifically comprises the following steps:
firstly, calculating the area of each connected region in the initial leakage segmentation image SRLR obtained in the step S7; then, the maximum value S of the communication area is calculated max (ii) a Finally, the area is less than 0.02S max The connected region is removed from the initial leakage segmentation image SRLR to obtain a final leakage segmentation image.
Referring to fig. 2, two types of fundus fluorescence angiography images applied in example 1 are shown.
Referring to fig. 3, the method of example 1 is used for comparing the results of leaky segmentation in clinical data sets with the results of expert manual segmentation. Table 1 below shows the accuracy index of the leakage segmentation result of the method of example 1 (based on the manual segmentation result of the expert).
TABLE 1
Image sequence number | Sensitivity | Specificity | Accuracy |
1 | 0.9905 | 0.9879 | 0.9880 |
2 | 0.8640 | 0.9895 | 0.9853 |
3 | 0.9012 | 0.9951 | 0.9897 |
4 | 0.8145 | 0.9991 | 0.9949 |
5 | 0.9311 | 0.9934 | 0.9907 |
6 | 1.0000 | 0.8773 | 0.8956 |
7 | 0.9939 | 0.9171 | 0.9253 |
8 | 1.0000 | 0.9173 | 0.9260 |
9 | 1.0000 | 0.9433 | 0.9501 |
10 | 1.0000 | 0.8702 | 0.8850 |
Mean | 0.9495 | 0.9490 | 0.9531 |
The qualitative comparison of fig. 3 illustrates the effectiveness of the method of example 1 in segmentation of the leaky region of a fluorescence contrast image of the fundus. As shown in table 1, the method of example 1 successfully segmented the leakage area of 10 samples with the expert segmentation results as a reference; specifically, the average sensitivity, specificity, and accuracy of the segmentation of the fundus fluorography image leakage region reached 0.9495, 0.9490, and 0.9531, respectively, and it was demonstrated that the method of example 1 can realize high-precision segmentation of a single-frame fundus fluorography image.
Example 2
The present embodiment provides a high-precision single-frame fundus fluorography image leakage region segmentation system that performs segmentation of a fundus fluorography image leakage region using the method of embodiment 1.
The present embodiment also provides a storage medium having stored thereon a computer program for implementing the method of embodiment 1 when executed.
The present embodiment also provides a computer device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the method of embodiment 1 when executing the computer program.
While embodiments of the invention have been disclosed above, it is not intended to be limited to the details shown in the description and the examples, which are set forth, but are fully applicable to various fields of endeavor as are suited to the particular use contemplated, and further modifications will readily occur to those skilled in the art, since the invention is not limited to the details shown and described without departing from the general concept as defined by the appended claims and their equivalents.
Claims (10)
1. A high-precision single-frame fundus fluorography image leakage area segmentation method is characterized by comprising the following steps:
s1, screening out a report of macular leakage or optic disc leakage from collected fundus fluorography image reports;
s2, selecting a later contrast period t from the reports screened in the step S1 1 To t 2 The minute fundus fluorography image is taken as a working data set;
s3, preprocessing the fundus fluorescence angiography image in the working data set: intercepting an image containing a macular leakage or optic disc leakage area to obtain a preprocessed fundus fluorography image I 0 ;
S4, utilizing the image I obtained in the step S3 0 Generating a retinal vascular mask;
s5, using the retinal blood vessel mask obtained in the step S4 to removeRemoving picture I 0 Obtaining a blood vessel-free image W, and carrying out Gaussian blur processing on the blood vessel-free image W to obtain an image W';
s6, performing desalination treatment on the background of the image W': removing the background of the image W 'to obtain a non-blood vessel image X, and then endowing the non-blood vessel image X with spatial Gaussian distribution weight to obtain an image X' after background desalination treatment;
s7, segmenting the image X' by adopting a method based on a Gaussian mixture model to obtain a primary leakage segmentation image;
and S8, optimizing the preliminary leakage segmentation image obtained in the step S7 to obtain a final leakage segmentation image.
2. The method for segmenting the leakage region of the high-precision single-frame fundus fluorography image according to claim 1, wherein the step S4 specifically comprises:
s4-1, utilizing Sobel operator to carry out comparison on image I 0 Processing to obtain image I 0 Is represented as:
wherein x, y represent pixels;
s4-2, processing the gradient image S (x, y) by adopting the erosion operation and the expansion operation respectively to obtain an eroded image Erode (S (x, y)) and an expanded image Diate (S (x, y)), and then determining that the region with the pixel difference value between the image Diate (S (x, y)) and the image Erode (S (x, y)) larger than P1 is a region S with blood vessels RBVR The calculation formula is as follows:
S RBVR ={(x,y)|Dilate(S(x,y))-Erode(S(x,y))>P1};
s4-3, for region S RBVR Performing a closing operation to fill the isolated region to obtain a retinal blood vessel mask S FBVR Expressed as:
S FBVR =Close(S RBVR )。
3. the method for segmenting the leakage region of the high-precision single-frame fundus fluorography image according to claim 2, wherein the step S5 specifically comprises:
s5-1, and for the fundus fluorography image I obtained in the step S3 0 And (3) performing opening operation processing to remove blood vessels U on the fundus fluorography image to obtain a blood vessel-free image with an affected leakage area, wherein U is expressed as:
U=Open(I);
s5-2, in the area covered by the retinal blood vessel mask, assigning the non-vascular image with the affected leakage area to the fundus fluorography image I 0 A non-vascular image W in which the leak region is not affected is obtained, and is expressed as:
s5-3, performing Gaussian blur processing on the non-vascular image W without the leakage area affected to fade the mask edge artifact and obtain a processed image W', which is expressed as:
W'=Gaussian Blur(W)。
4. the method for segmenting the leakage region of the high-precision single-frame fundus fluorescence angiography image according to claim 3, wherein the step S6 specifically comprises:
s6-1, carrying out corrosion operation on the non-blood vessel image W 'without the affected leakage area obtained in the step S5 to obtain a background image Erode (W') with the leakage area eliminated, and then subtracting the background image Erode (W ') from the image W' to obtain a non-blood vessel image X without the background, wherein the non-blood vessel image X is expressed as:
X=W'-Erode(W');
s6-2, endowing the background-removed non-blood vessel image with a spatial Gaussian distribution weight to obtain an image X' after background fading treatment, and expressing as follows:
X'(x,y)=X(x,y)×Gaussian Distribution(x,y)。
5. the method for segmenting the leakage region of the high-precision single-frame fundus fluorography image according to claim 4, wherein the step S7 specifically comprises:
s7-1, down-sampling the image X' obtained in the step S6 to obtain a down-sampled image D (X, y) which is expressed as:
D(x,y)=Downsample(X'(x,y));
s7-2, inputting the down-sampled image D (x, y) into a Gaussian mixture model and calculating the mean value M of the mu values of the image Gaussian mixture model:
M=Mean(Gaussian Mixture Model μ (D(x,y)));
s7-3, taking alpha times of the mean value M as a threshold, and carrying out threshold segmentation on the image X' obtained in the step S6 to obtain a primary leakage segmentation image S RLR Expressed as:
S RLR ={(x,y)|X'(x,y)>α×M};
wherein alpha is more than 0 and less than 1.
6. The method for segmenting the leakage region of the high-precision single-frame fundus fluorescence angiography image according to claim 5, wherein the step S8 specifically comprises:
first, the preliminary leakage segmentation image S obtained in step S7 is calculated RLR The area of each connected region; then, the maximum value S of the area of the communication region is calculated max (ii) a Finally, the area is less than beta S max Segmentation of the image S from the preliminary leakage RLR And (5) removing to obtain a final leakage segmentation image, wherein beta is more than 0 and less than 1.
7. The method for segmenting a leak region in a high-precision single-frame fundus fluorography image according to claim 6, wherein t is 1 =5,t 2 =6,P1=50,α=0.75,β=0.02。
8. A high-precision single-frame fundus fluorography image leakage region segmentation system, characterized in that it performs segmentation of fundus fluorography image leakage regions using the method according to any one of claims 1 to 7.
9. A storage medium on which a computer program is stored, characterized in that the program is adapted to carry out the method of any one of claims 1-7 when executed.
10. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the method according to any of claims 1-7 when executing the computer program.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211363937.6A CN115690124B (en) | 2022-11-02 | 2022-11-02 | High-precision single-frame fundus fluorescence contrast image leakage area segmentation method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211363937.6A CN115690124B (en) | 2022-11-02 | 2022-11-02 | High-precision single-frame fundus fluorescence contrast image leakage area segmentation method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115690124A true CN115690124A (en) | 2023-02-03 |
CN115690124B CN115690124B (en) | 2023-05-12 |
Family
ID=85048762
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211363937.6A Active CN115690124B (en) | 2022-11-02 | 2022-11-02 | High-precision single-frame fundus fluorescence contrast image leakage area segmentation method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115690124B (en) |
Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150248590A1 (en) * | 2014-03-03 | 2015-09-03 | Xerox Corporation | Method and apparatus for processing image of scene of interest |
US20160140724A1 (en) * | 2014-11-14 | 2016-05-19 | Huawei Technologies Co., Ltd. | Image processing method and apparatus |
CN109472781A (en) * | 2018-10-29 | 2019-03-15 | 电子科技大学 | A kind of diabetic retinopathy detection system based on serial structure segmentation |
WO2020122672A1 (en) * | 2018-12-13 | 2020-06-18 | 서울대학교병원 | Apparatus and method for automatically segmenting blood vessels by matching fp image and fag image |
US20200320686A1 (en) * | 2019-04-04 | 2020-10-08 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for processing fundus image |
US20200349712A1 (en) * | 2019-04-06 | 2020-11-05 | Kardiolytics Inc. | Method and system for machine learning based segmentation of contrast filled coronary artery vessels on medical images |
WO2021040258A1 (en) * | 2019-08-30 | 2021-03-04 | 국민대학교산학협력단 | Device and method for automatically diagnosing disease by using blood vessel segmentation in ophthalmic image |
CN112950737A (en) * | 2021-03-17 | 2021-06-11 | 中国科学院苏州生物医学工程技术研究所 | Fundus fluorescence radiography image generation method based on deep learning |
WO2021169128A1 (en) * | 2020-02-29 | 2021-09-02 | 平安科技(深圳)有限公司 | Method and apparatus for recognizing and quantifying fundus retina vessel, and device and storage medium |
US20210319556A1 (en) * | 2018-09-18 | 2021-10-14 | MacuJect Pty Ltd | Method and system for analysing images of a retina |
CN115100624A (en) * | 2022-05-27 | 2022-09-23 | 北京英泰智科技股份有限公司 | System and method for detecting zebra crossing in image |
-
2022
- 2022-11-02 CN CN202211363937.6A patent/CN115690124B/en active Active
Patent Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150248590A1 (en) * | 2014-03-03 | 2015-09-03 | Xerox Corporation | Method and apparatus for processing image of scene of interest |
US20160140724A1 (en) * | 2014-11-14 | 2016-05-19 | Huawei Technologies Co., Ltd. | Image processing method and apparatus |
US20210319556A1 (en) * | 2018-09-18 | 2021-10-14 | MacuJect Pty Ltd | Method and system for analysing images of a retina |
CN109472781A (en) * | 2018-10-29 | 2019-03-15 | 电子科技大学 | A kind of diabetic retinopathy detection system based on serial structure segmentation |
WO2020122672A1 (en) * | 2018-12-13 | 2020-06-18 | 서울대학교병원 | Apparatus and method for automatically segmenting blood vessels by matching fp image and fag image |
US20200320686A1 (en) * | 2019-04-04 | 2020-10-08 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for processing fundus image |
US20200349712A1 (en) * | 2019-04-06 | 2020-11-05 | Kardiolytics Inc. | Method and system for machine learning based segmentation of contrast filled coronary artery vessels on medical images |
WO2021040258A1 (en) * | 2019-08-30 | 2021-03-04 | 국민대학교산학협력단 | Device and method for automatically diagnosing disease by using blood vessel segmentation in ophthalmic image |
WO2021169128A1 (en) * | 2020-02-29 | 2021-09-02 | 平安科技(深圳)有限公司 | Method and apparatus for recognizing and quantifying fundus retina vessel, and device and storage medium |
CN112950737A (en) * | 2021-03-17 | 2021-06-11 | 中国科学院苏州生物医学工程技术研究所 | Fundus fluorescence radiography image generation method based on deep learning |
CN115100624A (en) * | 2022-05-27 | 2022-09-23 | 北京英泰智科技股份有限公司 | System and method for detecting zebra crossing in image |
Also Published As
Publication number | Publication date |
---|---|
CN115690124B (en) | 2023-05-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US5872861A (en) | Digital image processing method for automatic detection of stenoses | |
Sekhar et al. | Automated localisation of retinal optic disk using Hough transform | |
CN108765385B (en) | Double-source CT coronary artery automatic extraction method | |
CN110648338B (en) | Image segmentation method, readable storage medium, and image processing apparatus | |
Li et al. | A novel method of vessel segmentation for X-ray coronary angiography images | |
CN109087310B (en) | Meibomian gland texture region segmentation method and system, storage medium and intelligent terminal | |
CN111047607B (en) | Method for automatically segmenting coronary artery | |
CN110910404B (en) | Anti-noise data breast ultrasonic nodule segmentation method | |
SG172159A1 (en) | A method and system for determining the position of an optic cup boundary | |
Preethi et al. | Review of retinal blood vessel detection methods for automated diagnosis of Diabetic Retinopathy | |
CN109785320A (en) | It is a kind of to be classified to mammograms and known method for distinguishing using modified AlexNet model | |
Naveen et al. | Diabetic retinopathy detection using image processing | |
Akram et al. | Blood vessel enhancement and segmentation using wavelet transform | |
CN116188485A (en) | Image processing method, device, computer equipment and storage medium | |
Jana et al. | Liver fibrosis and nas scoring from ct images using self-supervised learning and texture encoding | |
CN115690124B (en) | High-precision single-frame fundus fluorescence contrast image leakage area segmentation method and system | |
Kumar et al. | Detection of microaneurysms and exudates from color fundus images by using SBGFRLS algorithm | |
Deng et al. | Automatic segmentation of ultrasound images of carotid atherosclerotic plaque based on Dense-UNet | |
Wen et al. | A novel lesion segmentation algorithm based on U-net network for tuberculosis CT image | |
Ashame et al. | Abnormality Detection in Eye Fundus Retina | |
Nageswari et al. | Automatic Detection and Classification of Diabetic Retinopathy using Modified UNET | |
Napier et al. | A CAD system for brain haemorrhage detection in head CT scans | |
Van Asperen et al. | Automatic artery/vein classification in 2D-DSA images of stroke patients | |
Akram et al. | Gabor wavelet based vessel segmentation in retinal images | |
Fitria et al. | Comparison of denoising methods applied to CTA images of 3D segmentation of aortic dissection |
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 |