CN103006175A - Method for positioning optic disk for eye fundus image on basis of PC (Phase Congruency) - Google Patents

Method for positioning optic disk for eye fundus image on basis of PC (Phase Congruency) Download PDF

Info

Publication number
CN103006175A
CN103006175A CN2012104616244A CN201210461624A CN103006175A CN 103006175 A CN103006175 A CN 103006175A CN 2012104616244 A CN2012104616244 A CN 2012104616244A CN 201210461624 A CN201210461624 A CN 201210461624A CN 103006175 A CN103006175 A CN 103006175A
Authority
CN
China
Prior art keywords
optic disc
eye fundus
image
fundus image
max
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
Application number
CN2012104616244A
Other languages
Chinese (zh)
Inventor
肖志涛
耿磊
张芳
吴骏
李筱涵
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tianjin Polytechnic University
Original Assignee
Tianjin Polytechnic University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Tianjin Polytechnic University filed Critical Tianjin Polytechnic University
Priority to CN2012104616244A priority Critical patent/CN103006175A/en
Publication of CN103006175A publication Critical patent/CN103006175A/en
Pending legal-status Critical Current

Links

Images

Landscapes

  • Eye Examination Apparatus (AREA)

Abstract

The invention belongs to the field of medical image treatment and relates to a method for positioning an optic disk for an eye fundus image on the basis of PC (Phase Congruency). According to the method, after four relatively prominent single-channel images for the optic disk are selected, a PC function is utilized to treat the selected images, and a logic AND is adopted for enhancing a result. The window scanning and the gray level accumulation are performed on the eye fundus image subjected to the enhancing treatment, so that the optic disk is positioned. The method is not influenced by the noise, luminance and contrast change in the images and is high in accuracy in positioning the optic disk for the eye fundus images of normal fundus, and slight and moderate as well as severe diabetes mellitus lesion.

Description

A kind of eye fundus image optic disc localization method based on phase equalization
Technical field
This patent relates to a kind of eye fundus image optic disc localization method based on phase equalization, the method can be determined the position of optic disc in eye fundus image, normal optical fundus and slight, moderate, severe diabetes pathological changes eye fundus image all there is very high accuracy, belong to technical field of image processing, can be applicable to Clinical Ophthalmology and with diagnosis and the treatment of retinopathy relevant disease.
Background technology
Medical Image Processing always is emphasis and the hot issue of studying in image processing and the analysis field with analysis.By the powerful measure of figure, image technique, quality and the display packing of medical image are greatly improved, so that treatment level improves greatly.Image processing techniques is introduced ophthalmology for many years, computational analysis by eye fundus image, the important fundus tissues such as optic disc, retinal vessel and macula lutea fovea centralis are carried out quantitative measurement, between normal and unusual, make clearly discriminating, can find early, exactly various eye pathological changes and considerable systemic disease, such as diabetes, hypertension, arteriosclerosis etc.Along with society to clinical optical fundus detect in the urgent need to, existing eye fundus image detection method shows many deficiencies in practical application and test, the difficulty of this field face and challenge also increase day by day.
Optic disc is one of the vital tissue on optical fundus, it is the pink circle that a diameter is about 1.5mm, be also referred to as papilla of optic nerve, it is the position of retinal nerve fiber and retinal vessel discrepancy eyeball, the neural impulse that reception produces from the visually-perceptible cell, and further conduct to brain formation vision by optic nerve.Its central authorities are funnel-shaped depression, are called physiological depression.Can see 4 arteria retinas and vein in the central authorities of optic disc, be the important guarantee of keeping retina nutrition.The parameters such as the shape of optic disc, area and the degree of depth are to weigh the important indicator of optical fundus health status, and the optic disc location is that ocular fundus image registration splicing, blood vessel segmentation, macula lutea and pathological changes are extracted and optic disc such as cuts apart at the basis that eye fundus image is processed.The multiformity such as imaging and individual differences causes the feature differences such as optic disc shape, color, size huge, and particularly the existence of various fundus pathological changes brings difficulty in the location automatically to optic disc.Work out and to satisfy the eye fundus image optic disc localization method that desired accuracy, objectivity, repeatable standard are detected in clinical optical fundus, for Clinical Ophthalmology research and with the diagnosis of retinopathy relevant disease with treat significant.
Up to now, the research work of relevant eye fundus image detection technique has been obtained many achievements, and Chinese scholars has proposed a lot of optic disc localization methods.Utilize the method for fuzzy clustering to set up relation between blood vessel and the optic disc such as: Hoover etc., and then obtain preferably locating effect (Hoover A, Goldbaum M.Locating the optic nerve in a retinal image using the fuzzy convergence of theblood vessels[J] .IEEE Transaction on Medical Imaging, 2003,22 (8): 951-958).Tobin etc. are based upon on the accurate extraction of vasoganglion, by analyzing its brightness, width and directional information location optic disc (Tobin K W, Chaum E, Govindasamy VP, et al.Detection of anatomic structures in human retinal imagery[J] .IEEE Transaction on Medical Imaging, 2007,26 (12): 1729-1739).Li Jupeng etc. by make up " class blood vessel " crossover network realize the optic disc location (Li Jupeng, Chen Houjin, Zhang Xinyuan. the optical fundus papilla of optic nerve based on crossover network is located automatically. electronics and information journal, 2009,31 (5): 1170-1174).Based on the different attribute of optical fundus optic disc, generally will have the optic disc localization method now and be divided into two kinds, based on the method for the feature of optic disc own with based on the method for retinal blood tubular construction.First kind method has the characteristic of higher brightness to locate optic disc based on optic disc circular configuration feature roughly with other position, relative optical fundus, for the retina that has pathological changes to occur, the robustness of this method is not strong, especially for the situation that fundus oculi disease occurs, the method can not obtain good positioning result.In addition, these class methods are removed optical fundus blood vessel by the morphology preprocess method mostly first, yet unified, the fixing Morphological Structuring Elements of the overall situation is used in the morphology pretreatment, and morphological transformation exists unintentional nonlinearity and irreversibility, therefore inevitably the optic disc edge feature is introduced larger change, cause the inaccurate of positioning result.The Equations of The Second Kind method is the attribute of retinal vessel networks converge point based on optic disc, need to be partitioned in advance blood vessel, extraction accuracy to vasoganglion has higher requirement, computation complexity is higher, and be only applicable to clearly retinal images of vasoganglion, be difficult to realize the optic disc location under the different images quality.
The eye fundus image background is complicated, and the factors such as the contrast of optic disc and background is low, uneven illumination even, optic disc size disunity all can affect positioning result, cause the inaccurate even track rejection in location.Therefore, need a kind of noise in the image, the impact that the brightness and contrast changes, eye fundus image optic disc localization method with versatility, robustness of not being subjected to.
Summary of the invention
For realizing that the optic disc positioning result is not subjected to the purpose of the factor affecting such as picture noise, brightness, the invention provides a kind of eye fundus image optic disc localization method based on phase equalization, may further comprise the steps: at first, be outstanding optic disc, eye fundus image is transformed to Lab, Yuv, Yiq and four color spaces of Hsv, select respectively the strongest channel image of optic disc and background contrasts in each space; Then, four selected channel image are carried out respectively phase equalization process, strengthen the result by logic "and" operation; At last, utilize the method location optic disc of window scanning and gray accumulation.
1. select to be fit to the color space passage of eye fundus image optic disc location: the inventive method has been selected a passage in Lab space, the u passage in Yuv space, the i passage in Yiq space and the s passage in Hsv space.Eye fundus image optic disc in these 4 passages is obviously outstanding and keep good local edge.
2. 4 single channel images selecting in the step 1 are carried out respectively PC and process, calculate the point of phase equalization maximum in the eye fundus image by estimating the peak value in the local energy function.This patent selects log Gabor wavelet function and its Hilbert conversion as the wave filter that calculates local energy.
3. the PC result of 4 single channel images carried out logic "and" operation, strengthen PC result.
4. select the image after 80 * 80 rectangular window scans whole logical “and”, the pixel mean-max in the calculation window utilizes the center-of-mass coordinate of window this moment just to locate optic disc.
5. in the rectangular window that step 4 obtains, carry out gray accumulation and calculate maximum gradation value X along X, Y-direction respectively Max, Y Max, the directions X gray threshold is made as 0.5*X Max, the Y-direction gray threshold is made as 0.5*Y Max, keep the image section greater than threshold value, calculate the centre coordinate of reserve part, finally determine the optic disc position with this coordinate.
Compare with existing method, beneficial effect of the present invention is:
1. phase information is not subjected to the impact of the factors such as contrast is low, uneven illumination is even, and therefore, this patent all has very strong robustness to the location of the optic disc under the different images quality.
2. this patent has been considered the skimble-scamble factor of Different Individual optic disc size, has adopted final definite optic disc position, two step location, has guaranteed the accuracy of location.
3. this patent can be realized automatically locating with slight, moderate, severe pathological changes eye fundus image optic disc at the bottom of the normal eyes, can realize the analytic function to patient's eye fundus image, for the diagnostic accuracy and the efficient that improve the doctor, improve the medical diagnosis level, enlarge ophthalmology numerical dialing system range of application important in inhibiting.
Description of drawings
Fig. 1: optic disc localization method flow chart of the present invention.
Fig. 2: the single channel eye fundus image in the different colours space.Fig. 2-1 is a passage eye fundus image in Lab space, and Fig. 2-2 is the u passage eye fundus image in Yuv space, and Fig. 2-3 is the i passage eye fundus image in Yiq space, and Fig. 2-4 is the s passage eye fundus image in Hsv space.
Fig. 3: after PC processes, the single channel eye fundus image in the different colours space.Fig. 3-1 is a passage eye fundus image in Lab space, and Fig. 3-2 is the u passage eye fundus image in Yuv space, and Fig. 3-3 is the i passage eye fundus image in Yiq space, and Fig. 3-4 is the s passage eye fundus image in Hsv space.
Fig. 4: image behind the logical “and”.
Fig. 5: clinical eye fundus image optic disc positioning experiment is figure as a result.First classifies original image as; Second classifies method positioning result of the present invention as.Fig. 5-1 is normal eye fundus image, and Fig. 5-2 is slight pathological changes eye fundus image, and Fig. 5-3 is moderate pathological changes eye fundus image, and Fig. 5-4 is severe pathological changes eye fundus image.
Fig. 6: utilize the STARE storehouse to carry out the optic disc positioning experiment, and with the contrast of the positioning result of Hoover method.First classifies original image as; Second classifies the positioning result of Hoover method as; The 3rd classifies the positioning result of the inventive method as.Fig. 6-1 is normal eye fundus image, and Fig. 6-2~6-5 is the pathological changes eye fundus image.
The specific embodiment
Flow chart of the present invention at first for outstanding optic disc, transforms to Lab, Yuv, Yiq, four color spaces of Hsv with original image as shown in Figure 1; In each space, select a single channel image to carry out PC and process, by logic "and" operation enhancement process result; Utilize the method location optic disc of window scanning and gray accumulation.Below in conjunction with accompanying drawing, the specific implementation process of technical solution of the present invention is illustrated.
1. the color space passage is chosen:
Select a passage, the u passage in Yuv space, the i passage in Yiq space and the s passage in Hsv space in Lab space.As shown in Figure 2, original image transformed to these 4 passages after, optic disc is obviously outstanding and keep good local edge, is beneficial to follow-up optic disc location.
2. based on phase equalization eye fundus image is processed:
The phase equalization function definition is as follows:
PC ( x ) = max φ ‾ ( x ) ∈ [ 0,2 π ] Σ n A n cos ( φ n ( x ) - φ ‾ ( x ) ) Σ n A n
In the formula, A nBe the amplitude of nth harmonic cosine component, φ nThe phase place of n secondary frequency components,
Figure BSA00000806419100032
Weighted average for the local phase of all Fourier's items of being positioned at this point.
The point of phase equalization maximum can equivalence be the peak value in the local energy function.Local energy can be estimated by following formula:
LE = | | E ( x ) | | = F 2 + H 2 = ( I ( x ) * M even ) 2 + ( I ( x ) * M odd ) 2
For one-dimensional signal F (x), local energy is defined as square root sum square of signal F (x) and its Hilbert conversion H (x).Two parts of local energy can be estimated by the convolution of signal I (x) and pair of orthogonal wave filter, and a wave filter is even symmetry, is M Even, another is odd symmetric, is M Odd
This patent selects log Gabor wavelet function and its Hilbert conversion as the wave filter that calculates local energy.On the linear frequency yardstick, the form of the transfer function of log Gabor function is:
g ( ω ) = e - ( log ( ω / ω 0 ) ) 2 2 ( log ( β / ω 0 ) ) 2
Here ω 0Mid frequency for wave filter.For guaranteeing the constant shape of wave filter, for different mid frequency ω 0, β/ω 0Must be consistent.
4 single channel images selecting are previously carried out respectively PC process, the result as shown in Figure 3.
3. logic "and" operation:
PC result to 4 single channel images carries out logic "and" operation, and the result as shown in Figure 4.
4. the eye fundus image behind the logical “and” is adopted final definite optic disc position, two step location:
4.1 first location: select the image after 80 * 80 rectangular window scans whole logical “and”, the pixel average in the calculation window, the zone of pixel average maximum namely is the position of optic disc, utilizes the center-of-mass coordinate location optic disc of window this moment.
4.2 accurately locate: in rectangular window obtained in the previous step, carry out gray accumulation and calculate maximum gradation value X along X, Y-direction respectively Max, Y Max, the directions X gray threshold is made as 0.5*X Max, the Y-direction gray threshold is made as 0.5*Y Max, keep the image section greater than threshold value, calculate the centre coordinate of reserve part, finally determine the optic disc position with this coordinate.
Adopt this patent method that the eye fundus image in clinical collection and STARE, the DRIVE storehouse is tested, the result is shown in Fig. 5, Fig. 6 and table 1.Method of the present invention is positioned with good effect to the eye fundus image optic disc, can to normal optical fundus and in various degree pathological changes optical fundus standard to the location optic disc, have very strong robustness.
Table 1: utilize the DRIVE storehouse to carry out the optic disc positioning experiment, and with the contrast of the position success rate of Tobin, Hoover method.
Figure BSA00000806419100041

Claims (4)

1. the eye fundus image optic disc localization method based on phase equalization comprises the following steps:
(1) the space single channel image that gets colors;
(2) utilize the phase equalization function that selected single channel eye fundus image is processed;
(3) the PC result of single channel image carried out logic "and" operation;
(4) select image after 80 * 80 rectangular window scans whole logical “and”, the pixel mean-max in the calculation window utilizes the center-of-mass coordinate of window this moment just to locate optic disc;
(5) in the rectangular window that step (4) obtains, carry out gray accumulation and calculate maximum gradation value X along Y-direction not respectively Max, Y Max, the directions X gray threshold is made as 0.5*X Max, the Y-direction gray threshold is made as 0.5*Y Max, keep the image section greater than threshold value, calculate the centre coordinate of reserve part, with the final accurately location of this coordinate optic disc.
2. eye fundus image optic disc localization method according to claim 1, it is characterized in that, in the step (1), only choose the color space single channel image that is fit to eye fundus image optic disc location, image is not carried out any morphology pretreatment, guaranteed the accuracy of optic disc edge feature.
3. eye fundus image optic disc localization method according to claim 1, it is characterized in that, in the step (2), use phase equalization sign optic disc feature, need not waveform is carried out any hypothesis, just in Fourier transform domain, seek characteristic point by the consistent order of phase place.
4. eye fundus image optic disc localization method according to claim 1 is characterized in that, in step (4), (5), uses elder generation just to locate pinpoint twice localization method again and determines the optic disc position, improves accurate positioning.
CN2012104616244A 2012-11-14 2012-11-14 Method for positioning optic disk for eye fundus image on basis of PC (Phase Congruency) Pending CN103006175A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2012104616244A CN103006175A (en) 2012-11-14 2012-11-14 Method for positioning optic disk for eye fundus image on basis of PC (Phase Congruency)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2012104616244A CN103006175A (en) 2012-11-14 2012-11-14 Method for positioning optic disk for eye fundus image on basis of PC (Phase Congruency)

Publications (1)

Publication Number Publication Date
CN103006175A true CN103006175A (en) 2013-04-03

Family

ID=47955675

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2012104616244A Pending CN103006175A (en) 2012-11-14 2012-11-14 Method for positioning optic disk for eye fundus image on basis of PC (Phase Congruency)

Country Status (1)

Country Link
CN (1) CN103006175A (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104835150A (en) * 2015-04-23 2015-08-12 深圳大学 Learning-based eyeground blood vessel geometric key point image processing method and apparatus
CN108364285A (en) * 2018-01-24 2018-08-03 首都师范大学 A kind of crack extract methods and applications based on iterative phase consistency
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
CN110009627A (en) * 2019-04-11 2019-07-12 北京百度网讯科技有限公司 Method and apparatus for handling image
CN110874597A (en) * 2018-08-31 2020-03-10 福州依影健康科技有限公司 Blood vessel feature extraction method, device and system for fundus image and storage medium
CN113269737A (en) * 2021-05-17 2021-08-17 西安交通大学 Method and system for calculating diameter of artery and vein of retina

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2005261799A (en) * 2004-03-22 2005-09-29 Canon Inc Ophthalmologic image processor
RU2305484C1 (en) * 2006-01-18 2007-09-10 Федеральное государственное учреждение "Межотраслевой научно-технический комплекс "Микрохирургия глаза" имени академика С.Н. Федорова Федерального агентства по здравоохранению и социальному развитию" Differential diagnosis method for determining stagnated optic nerve disk and optic nerve neuritis
WO2010073655A1 (en) * 2008-12-26 2010-07-01 Canon Kabushiki Kaisha Imaging apparatus and method for taking image of eyeground by optical coherence tomography

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2005261799A (en) * 2004-03-22 2005-09-29 Canon Inc Ophthalmologic image processor
RU2305484C1 (en) * 2006-01-18 2007-09-10 Федеральное государственное учреждение "Межотраслевой научно-технический комплекс "Микрохирургия глаза" имени академика С.Н. Федорова Федерального агентства по здравоохранению и социальному развитию" Differential diagnosis method for determining stagnated optic nerve disk and optic nerve neuritis
WO2010073655A1 (en) * 2008-12-26 2010-07-01 Canon Kabushiki Kaisha Imaging apparatus and method for taking image of eyeground by optical coherence tomography

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
M.ASHRAFUL AMIN AL: "phase congruency based retinal vessel segmentation", 《PROCEEDINGS OF THE EIGHTH INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND CYBERNETICS》 *

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104835150A (en) * 2015-04-23 2015-08-12 深圳大学 Learning-based eyeground blood vessel geometric key point image processing method and apparatus
CN104835150B (en) * 2015-04-23 2018-06-19 深圳大学 A kind of optical fundus blood vessel geometry key point image processing method and device based on study
CN108364285A (en) * 2018-01-24 2018-08-03 首都师范大学 A kind of crack extract methods and applications based on iterative phase consistency
CN108364285B (en) * 2018-01-24 2021-06-18 首都师范大学 Iterative phase consistency-based crack extraction method and application
CN110874597A (en) * 2018-08-31 2020-03-10 福州依影健康科技有限公司 Blood vessel feature extraction method, device and system for fundus image and storage medium
CN110874597B (en) * 2018-08-31 2023-03-24 福州依影健康科技有限公司 Blood vessel feature extraction method, device and system for fundus image and storage medium
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
CN109472781B (en) * 2018-10-29 2022-02-11 电子科技大学 Diabetic retinopathy detection system based on serial structure segmentation
CN110009627A (en) * 2019-04-11 2019-07-12 北京百度网讯科技有限公司 Method and apparatus for handling image
CN110009627B (en) * 2019-04-11 2021-06-18 北京康夫子健康技术有限公司 Method and apparatus for processing image
CN113269737A (en) * 2021-05-17 2021-08-17 西安交通大学 Method and system for calculating diameter of artery and vein of retina
CN113269737B (en) * 2021-05-17 2024-03-19 北京鹰瞳科技发展股份有限公司 Fundus retina artery and vein vessel diameter calculation method and system

Similar Documents

Publication Publication Date Title
CN107529646B (en) Non-contact heart rate measurement method and device based on Euler image amplification
Sarathi et al. Blood vessel inpainting based technique for efficient localization and segmentation of optic disc in digital fundus images
CN103006175A (en) Method for positioning optic disk for eye fundus image on basis of PC (Phase Congruency)
CN105513077B (en) A kind of system for diabetic retinopathy screening
Tavakoli et al. A complementary method for automated detection of microaneurysms in fluorescein angiography fundus images to assess diabetic retinopathy
Rangayyan et al. Detection of the optic nerve head in fundus images of the retina with Gabor filters and phase portrait analysis
CN110269600A (en) Based on polynary empirical mode decomposition and the contactless video heart rate detection method for combining blind source separating
Agarwal et al. Automatic glaucoma detection using adaptive threshold based technique in fundus image
CN104835150B (en) A kind of optical fundus blood vessel geometry key point image processing method and device based on study
Tam et al. Speed quantification and tracking of moving objects in adaptive optics scanning laser ophthalmoscopy
Abdel-Ghafar et al. Progress towards automated detection and characterization of the optic disc in glaucoma and diabetic retinopathy
CN105678780A (en) Video heart rate detection method removing interference of ambient light variation
CN107374583A (en) One kind eliminates pseudo- image method and storage medium and imaging system in OCT blood vessel imagings
CN101999885A (en) Endogenous optical imaging method for automatically separating arteries and veins
CN107292835A (en) A kind of method and device of eye fundus image retinal vessel Automatic Vector
CN109949277A (en) A kind of OCT image quality evaluating method based on sequence study and simplified residual error network
Zhu et al. Digital image processing for ophthalmology: Detection of the optic nerve head
SG172159A1 (en) A method and system for determining the position of an optic cup boundary
Kirsch et al. Automated localization of magnetoencephalographic interictal spikes by adaptive spatial filtering
Kumari et al. Blood vessel extraction using wiener filter and morphological operation
Shabbir et al. Automatic cotton wool spots extraction in retinal images using texture segmentation and Gabor wavelet
CN108665474B (en) B-COSFIRE-based retinal vessel segmentation method for fundus image
Karthika et al. Retinal image analysis using contourlet transform and multistructure elements morphology by reconstruction
Mahfouz et al. Ultrafast localization of the optic disc using dimensionality reduction of the search space
Zardadi et al. Unsupervised segmentation of retinal blood vessels using the human visual system line detection model

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C02 Deemed withdrawal of patent application after publication (patent law 2001)
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20130403