CN110599417A - Retina image blood vessel extraction method based on Hessian matrix - Google Patents
Retina image blood vessel extraction method based on Hessian matrix Download PDFInfo
- Publication number
- CN110599417A CN110599417A CN201910832201.0A CN201910832201A CN110599417A CN 110599417 A CN110599417 A CN 110599417A CN 201910832201 A CN201910832201 A CN 201910832201A CN 110599417 A CN110599417 A CN 110599417A
- Authority
- CN
- China
- Prior art keywords
- blood vessel
- image
- hessian matrix
- extraction method
- vessel extraction
- 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.)
- Pending
Links
- 210000004204 blood vessel Anatomy 0.000 title claims abstract description 40
- 239000011159 matrix material Substances 0.000 title claims abstract description 29
- 238000000605 extraction Methods 0.000 title claims abstract description 18
- 210000001525 retina Anatomy 0.000 title claims abstract description 10
- 238000012545 processing Methods 0.000 claims abstract description 22
- 230000004256 retinal image Effects 0.000 claims abstract description 14
- 238000000034 method Methods 0.000 claims description 22
- 238000001514 detection method Methods 0.000 abstract description 7
- 230000000694 effects Effects 0.000 description 4
- 206010012689 Diabetic retinopathy Diseases 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 2
- 201000010099 disease Diseases 0.000 description 2
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 201000004569 Blindness Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 206010012601 diabetes mellitus Diseases 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000036039 immunity Effects 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000002207 retinal effect Effects 0.000 description 1
- 230000004393 visual impairment Effects 0.000 description 1
Classifications
-
- G06T5/73—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- 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
- G06T2207/30101—Blood vessel; Artery; Vein; Vascular
Abstract
The invention discloses a retina image blood vessel extraction method based on a Hessian matrix, which relates to the field of image processing and comprises the steps of firstly, obtaining a retina G channel image; then, carrying out three times of Hessian matrix processing on the obtained image by using a filter; then carrying out binarization processing on the obtained image; then, the obtained image is repaired by using closed operation; and finally, denoising and eliminating the obtained image by using a connected domain to obtain enhanced noise due to multiple Hessian matrixes. The invention takes opencv as a platform to write programs, combines the related knowledge of the Hessian matrix, finishes the extraction of blood vessels in the retinal image, and provides a scheme of 'three times Hessian matrix processing and connected domain denoising', greatly improves the precision of blood vessel detection and the noise-resistant capability of the scheme, and obtains good experimental results.
Description
Technical Field
The invention relates to the field of image processing, in particular to a retina image blood vessel extraction method based on a Hessian matrix.
Background
Diabetic retinopathy is a complication caused by diabetes mellitus, and has become the leading cause of blindness among people working in developed countries. Although diabetic retinopathy is still not completely cured under current medical conditions, early discovery and hospitalization still minimizes visual loss. At present, diabetic retinopathy is mainly used for disease diagnosis and clinical research through the characteristics of blood vessel diameter, color, curvature and the like of retinal images, and the blood vessels have the following four characteristics: (1) the gray value distribution of the cross section of the blood vessel is approximate to a Gaussian curve; (2) the blood vessel can be seen as a continuous number of line segments; (3) the direction of the blood vessels and the gray values are continuously varied rather than abrupt; (4) the blood vessels communicate with each other in a tree-like structure and start from the optic disc. Current blood vessel detection based on the above features can be roughly classified into four categories: a matched filter method; a mathematical morphological method; a blood vessel tracking method; a classifier method. However, these methods can hardly reduce noise well and maintain more detailed parts of the image, i.e. many fine blood vessels in the image, and the detection of these lesions is of great significance for early disease diagnosis.
Therefore, those skilled in the art are dedicated to developing a retinal image blood vessel extraction method based on a Hessian matrix, which can achieve better effects in both denoising and detail extraction in the process of denoising and extracting a blood vessel image.
Disclosure of Invention
In view of the above-mentioned drawbacks of the prior art, the technical problem to be solved by the present invention is how to retain more details, i.e. how to improve the accuracy of blood vessel detection and the noise immunity of the solution, while performing noise reduction processing.
In order to achieve the above object, the present invention provides a retinal image blood vessel extraction method based on a Hessian matrix, which is characterized by comprising the following steps:
step 1, obtaining a retina G channel image;
step 2, performing three times of Hessian matrix processing on the image obtained in the step 1 by using a filter;
step 3, carrying out binarization processing on the image obtained in the step 2;
step 4, repairing the image obtained in the step 3 by using closed operation;
and 5, denoising the image obtained in the step 4 by using a connected domain.
Further, the step 2 further comprises:
step 2.1, setting corresponding parameters, and performing first processing on the image obtained in the step 1 by using a Hessian matrix method;
step 2.2, inverting the gray value of the image obtained in the step 2.1;
2.3, setting corresponding parameters, and performing secondary processing on the image obtained in the step 2.2 by using a Hessian matrix method;
step 2.4, inverting the gray value of the image obtained in the step 2.3;
and 2.5, setting corresponding parameters, and performing third processing on the image obtained in the step 2.4 by using a Hessian matrix method.
Further, the filter in step 1 is a multi-scale linear filter.
Further, in step 3, binarization is performed by an OTSU algorithm.
Further, the formulas for inverting the gray values in step 2.2 and step 2.4 are both:
I(x,y)＝255-I′(x,y)
wherein: i (x, y) is the gray scale value of a certain point after inversion, and I' (x, y) is the gray scale value of a certain point in the original image.
Preferably, the scale factor in the corresponding parameters set in step 2.1 is [1.2,3.2], and the step size is 0.5.
Preferably, the strongest filter output coefficient of said corresponding parameters set in step 2.1 is 3.5.
Preferably, the scale factor in the corresponding parameters set in step 2.3 is [1.2,3.2], and the step size is 0.5.
Preferably, the scale factor in the corresponding parameters set in step 2.5 is [1.2,3.2], and the step size is 0.5.
Preferably, the strongest filter output coefficients of the respective parameters set in step 2.3 and step 2.5 are both 0.3.
The invention has the following beneficial technical effects:
1. compared with the method using the Hessian matrix for one time, the method using the three times Hessian matrix for blood vessel enhancement improves the contrast between the blood vessel at the position of the thin blood vessel and the background, so that more blood vessel details can be obtained after binarization, and the precision of blood vessel detection is improved.
2. The noise generated by Hessian matrix enhancement is removed from the image subjected to connected domain denoising, and meanwhile, the details of the image, especially a plurality of tiny blood vessels, are well kept, and the noise resistance of blood vessel detection is improved.
The conception, the specific structure and the technical effects of the present invention will be further described with reference to the accompanying drawings to fully understand the objects, the features and the effects of the present invention.
Drawings
FIG. 1 is a flow chart of a method in accordance with a preferred embodiment of the present invention;
FIG. 2 is a retinal fundus image G channel image of a preferred embodiment of the present invention;
FIG. 3 is a once Hessian matrix enhanced image of a preferred embodiment of the present invention;
FIG. 4 is a cubic Hessian matrix enhanced image of a preferred embodiment of the present invention;
FIG. 5 is a noisy image in accordance with a preferred embodiment of the present invention;
FIG. 6 is a diagram illustrating a connected component denoised image according to a preferred embodiment of the present invention.
Detailed Description
The preferred embodiments of the present invention will be described below with reference to the accompanying drawings for clarity and understanding of technical contents. The present invention may be embodied in many different forms of embodiments and the scope of the invention is not limited to the embodiments set forth herein.
The Hessian method is a method for extracting the characteristic direction of an image by high-order differentiation. The Hessian method considers that the direction of the feature vector having the largest mode is perpendicular to the image feature direction, and the direction perpendicular thereto is considered to be parallel to the image feature direction. For a linear model constructed by a gaussian function, it can be represented by the second derivative with the maximum absolute value orthogonal to the straight line and the second derivative with the small absolute value along the line direction, which is exactly the collective meaning represented by the two eigenvalues of the two-digit Hessian matrix. The property of the Hessian matrix is applied to blood vessel detection, and the noise in the retina image is removed by designing a linear enhancement filter function, so that the retina blood vessel is detected and enhanced.
FIG. 1 is a flowchart of a method of a preferred embodiment of the present invention, including start-;
step 1, obtaining a retina G channel image, wherein the obtained image is shown in figure 2;
step 2, performing three times of Hessian matrix processing on the image obtained in the step 1 by using a multi-scale linear filter:
(1) performing first processing on the image obtained in the step 1 by using a Hessian matrix method, setting corresponding parameter scale factors to be [1.2 and 3.2], setting the step length to be 0.5, and setting the strongest filtering output coefficient to be 3.5, wherein the processed image is shown in FIG. 3;
(2) inverting the gray value of the image obtained in the step 2.1, wherein the formula is that I (x, y) is 255-I '(x, y), where I (x, y) is the gray value of a certain point after inversion, and I' (x, y) is the gray value of a certain point in the original image;
(3) performing secondary processing on the image obtained in the step 2.2 by using a Hessian matrix method, and setting corresponding parameter scale factors to be [1.2 and 3.2], the step length to be 0.5 and the strongest filter output coefficient to be 0.3;
(4) carrying out gray value inversion on the image obtained in the step 2.3;
(5) and performing third processing on the image obtained in the step 2.4 by using a Hessian matrix method, setting corresponding parameter scale factors to be [1.2 and 3.2], setting the step length to be 0.5, and setting the strongest filtering output coefficient to be 0.3, wherein the processed image is shown in FIG. 4.
Step 3, performing binarization processing on the image obtained in the step 2 by using an OTSU algorithm;
step 4, because the blood vessel fracture will occur at the position of the thin blood vessel on the image after the binarization processing is carried out in the step 3, the image obtained in the step 3 is repaired by using a closing operation, and the image processed in the step is as shown in fig. 5;
and 5, denoising the image obtained in the step 4 by using a connected domain, so that noise enhanced by a plurality of Hessian matrixes can be eliminated, and the processed image is shown in FIG. 6.
The foregoing detailed description of the preferred embodiments of the invention has been presented. It should be understood that numerous modifications and variations could be devised by those skilled in the art in light of the present teachings without departing from the inventive concepts. Therefore, the technical solutions available to those skilled in the art through logic analysis, reasoning and limited experiments based on the prior art according to the concept of the present invention should be within the scope of protection defined by the claims.
Claims (10)
1. A retinal image blood vessel extraction method based on a Hessian matrix is characterized by comprising the following steps:
step 1, obtaining a retina G channel image;
step 2, performing three times of Hessian matrix processing on the image obtained in the step 1 by using a filter;
step 3, carrying out binarization processing on the image obtained in the step 2;
step 4, repairing the image obtained in the step 3 by using closed operation;
and 5, denoising the image obtained in the step 4 by using a connected domain.
2. The retinal image blood vessel extraction method according to claim 1, wherein the step 2 further comprises:
step 2.1, setting corresponding parameters, and performing first processing on the image obtained in the step 1 by using a Hessian matrix method;
step 2.2, inverting the gray value of the image obtained in the step 2.1;
2.3, setting corresponding parameters, and performing secondary processing on the image obtained in the step 2.2 by using a Hessian matrix method;
step 2.4, inverting the gray value of the image obtained in the step 2.3;
and 2.5, setting corresponding parameters, and performing third processing on the image obtained in the step 2.4 by using a Hessian matrix method.
3. The retinal image blood vessel extraction method according to claim 1, wherein the filter in step 1 is a multi-scale linear filter.
4. The retinal image blood vessel extraction method according to claim 1, characterized in that in step 3, binarization is performed by OTSU algorithm.
5. The retinal image blood vessel extraction method according to claim 2, wherein the formula for inverting the gray value in step 2.2 and step 2.4 is:
I(x,y)＝255-I′(x,y)
wherein: i (x, y) is the gray scale value of a certain point after inversion, and I' (x, y) is the gray scale value of a certain point in the original image.
6. The retinal image blood vessel extraction method according to claim 2, wherein the scale factor in the corresponding parameter set in step 2.1 is [1.2,3.2], and the step size is 0.5.
7. A retinal image blood vessel extraction method according to claim 6, characterized in that the strongest filter output coefficient of the respective parameters set in step 2.1 is 3.5.
8. The retinal image blood vessel extraction method according to claim 2, wherein the scale factor in the corresponding parameter set in step 2.3 is [1.2,3.2], and the step size is 0.5.
9. The retinal image blood vessel extraction method according to claim 2, wherein the scale factor in the corresponding parameter set in step 2.5 is [1.2,3.2], and the step size is 0.5.
10. A retinal image blood vessel extraction method according to claims 8-9, characterized in that the strongest filter output coefficients of the respective parameters set in step 2.3 and step 2.5 are both 0.3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910832201.0A CN110599417A (en) | 2019-09-04 | 2019-09-04 | Retina image blood vessel extraction method based on Hessian matrix |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910832201.0A CN110599417A (en) | 2019-09-04 | 2019-09-04 | Retina image blood vessel extraction method based on Hessian matrix |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110599417A true CN110599417A (en) | 2019-12-20 |
Family
ID=68857528
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910832201.0A Pending CN110599417A (en) | 2019-09-04 | 2019-09-04 | Retina image blood vessel extraction method based on Hessian matrix |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110599417A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112862850A (en) * | 2021-02-03 | 2021-05-28 | 北京阅影科技有限公司 | Digital silhouette angiography image processing method and device |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104835157A (en) * | 2015-05-04 | 2015-08-12 | 北京工业大学 | Eye fundus image optical cup automatic segmentation method based on improved PDE image repairing |
| CN106296664A (en) * | 2016-07-30 | 2017-01-04 | 上海联影医疗科技有限公司 | Vessel extraction method |
| CN106407917A (en) * | 2016-09-05 | 2017-02-15 | 山东大学 | Dynamic scale distribution-based retinal vessel extraction method and system |
| CN106815853A (en) * | 2016-12-14 | 2017-06-09 | 海纳医信（北京）软件科技有限责任公司 | To the dividing method and device of retinal vessel in eye fundus image |
| CN107871322A (en) * | 2016-09-27 | 2018-04-03 | 北京眼神科技有限公司 | Iris segmentation method and apparatus |
| CN108109159A (en) * | 2017-12-21 | 2018-06-01 | 东北大学 | It is a kind of to increase the retinal vessel segmenting system being combined based on hessian matrixes and region |
-
2019
- 2019-09-04 CN CN201910832201.0A patent/CN110599417A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104835157A (en) * | 2015-05-04 | 2015-08-12 | 北京工业大学 | Eye fundus image optical cup automatic segmentation method based on improved PDE image repairing |
| CN106296664A (en) * | 2016-07-30 | 2017-01-04 | 上海联影医疗科技有限公司 | Vessel extraction method |
| CN106407917A (en) * | 2016-09-05 | 2017-02-15 | 山东大学 | Dynamic scale distribution-based retinal vessel extraction method and system |
| CN107871322A (en) * | 2016-09-27 | 2018-04-03 | 北京眼神科技有限公司 | Iris segmentation method and apparatus |
| CN106815853A (en) * | 2016-12-14 | 2017-06-09 | 海纳医信（北京）软件科技有限责任公司 | To the dividing method and device of retinal vessel in eye fundus image |
| CN108109159A (en) * | 2017-12-21 | 2018-06-01 | 东北大学 | It is a kind of to increase the retinal vessel segmenting system being combined based on hessian matrixes and region |
Non-Patent Citations (1)
| Title |
|---|
| 蒋先刚等: "基于Hessian特征的视网膜血管图像的增强滤波算法", 《华东交通大学学报》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112862850A (en) * | 2021-02-03 | 2021-05-28 | 北京阅影科技有限公司 | Digital silhouette angiography image processing method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108986106B (en) | Automatic segmentation method for retinal blood vessels for glaucoma | |
| Tjandrasa et al. | Optic nerve head segmentation using hough transform and active contours | |
| WO2010138645A2 (en) | Blood vessel segmentation with three-dimensional spectral domain optical coherence tomography | |
| CN108109159B (en) | Retina blood vessel segmentation system based on hessian matrix and region growing combination | |
| CN113470137B (en) | IVOCT image guide wire artifact removing method based on gray-scale weighting | |
| CN109919954B (en) | Target object identification method and device | |
| Washburn | Investigation of severity level of diabetic retinopathy using adaboost classifier algorithm | |
| Elbalaoui et al. | Automatic detection of blood vessel in retinal images | |
| CN109087310B (en) | Meibomian gland texture region segmentation method and system, storage medium and intelligent terminal | |
| Andayani et al. | Identification of retinoblastoma using backpropagation neural network | |
| CN115205315A (en) | Fundus image enhancement method for maintaining ophthalmologic physical signs | |
| CN110599417A (en) | Retina image blood vessel extraction method based on Hessian matrix | |
| Afonso et al. | Hair detection in dermoscopic images using percolation | |
| Dash et al. | Retinal blood vessels extraction from fundus images using an automated method | |
| Shabbir et al. | A comparison and evaluation of computerized methods for blood vessel enhancement and segmentation in retinal images | |
| Mirajkar et al. | Feature extraction of diabetic retinopathy images | |
| Desiani et al. | A robust techniques of enhancement and segmentation blood vessels in retinal image using deep learning | |
| Du et al. | Retinal image segmentation based on Mumford-Shah model and Gabor wavelet filter | |
| CN108154512A (en) | It is a kind of based on the multiple retinal images blood vessel segmentation system for going trend analysis | |
| Ozgunalp et al. | Semantic segmentation of retinal vessels using SegNet | |
| Kabir | A rule based segmentation approaches to extract retinal blood vessels in fundus image | |
| Safmi et al. | The Augmentation Data of Retina Image for Blood Vessel Segmentation Using U-Net Convolutional Neural Network Method. | |
| Halim et al. | Enhancement of automatic classification of arcus senilis-nonarcus senilis using convolutional neural network | |
| CN112767375B (en) | OCT image classification method, system and equipment based on computer vision characteristics | |
| Abdulsahib et al. | Automated Approach for Extraction of Retinal Blood Vessels |
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 |