CN106214120A - A kind of methods for screening of glaucoma - Google Patents
A kind of methods for screening of glaucoma Download PDFInfo
- Publication number
- CN106214120A CN106214120A CN201610689336.2A CN201610689336A CN106214120A CN 106214120 A CN106214120 A CN 106214120A CN 201610689336 A CN201610689336 A CN 201610689336A CN 106214120 A CN106214120 A CN 106214120A
- Authority
- CN
- China
- Prior art keywords
- glaucoma
- cup
- optic disc
- optic
- image
- 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
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
- A61B3/0016—Operational features thereof
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
- A61B3/10—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions
- A61B3/12—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions for looking at the eye fundus, e.g. ophthalmoscopes
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
- A61B3/10—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions
- A61B3/16—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions for measuring intraocular pressure, e.g. tonometers
- A61B3/165—Non-contacting tonometers
Abstract
The present invention is the methods for screening of a kind of glaucoma, utilize computer that fundus photograph is carried out automatical analysis, quickly calculate the cup disc ratio C/D ratio of optic cup and optic disc in real time, it is characterised in that the glaucoma value-at-risk contrast of cup disc ratio with setting is judged whether the probability of glaucoma.Advantages of the present invention: produce cup disc ratio with optic nerve Progressive symmetric erythrokeratodermia damage in glaucoma pathological changes and be changed to basis, it is achieved under area of computer aided, full-automatic fundus photograph diagosis is analyzed.Due to image blurring process, OTSU algorithm and the introducing of Computing, compared with the method in the past utilizing merely fundus photography diagosis to diagnose, substantially increase accuracy and real-time that result judges, evaded detection equipment costly simultaneously.By the examination of low-cost high-efficiency, it is achieved find the morning of glaucoma, early diagnose；Big data expansion development system under gradual perfection three cascade network system by the way of the foundation of data base and step-sizing are uploaded step by step.
Description
Technical field
The methods for screening of a kind of glaucoma of the present invention, relates to medical image processing technology field.
Technical background
Glaucoma is one of the whole world three big blinding oculopathy, is global important public hygiene problem, and long-term optic nerve damages
Wound causes irreversibility blind the most at last, all causes heavy pressure and burden for sufferers themselves, family and society.
China is the ill big country of glaucoma, and preventing and treating glaucoma shoulders heavy responsibilities.Money according to the World Health Organization (WHO)
Material, the Endothelium in Patients with Primary Glaucoma number about 15,780,000 (1111 ~ 23,640,000) of China in 2010, wherein primary open-angle
Glaucoma (POAG) 8,310,000 (6,700,000 ~ 10,420,000), accounts for the 18.6% of whole world POAG；Primary angle-closure glaucoma
(PACG) 7,470,000 (4,420,000 ~ 13,220,000), account for the 47.5% of whole world PACG.More than the 40 years old (total population 5.93 in China
Hundred million), in crowd, prevalence of glaucoma is 2.66%(1.87% ~ 3.98%), POAG prevalence 1.4%, PACG prevalence
1.26%.Estimating 2020, China's primary glaucoma is up to about 21,820,000 (11556 ~ 32,010,000), wherein 40 years old with
Upper (total population about 7.15 hundred million at that time) prevalence is 3.05%(1.64% ~ 1.41%).
Early discovery glaucoma also carries out rational therapy and can avoid blinding.As Fig. 1 glaucoma early intervention was done with late period
Shown in the difference of pre-final result, primary glaucoma belongs to chronic progressive disease, carries out intraocular pressure lowering treatment and make after early discovery
Reach target intraocular pressure, the state of an illness of most patient can be made can be controlled by the method such as medicine and surgical laser, guarantor
Hold existing visual function and no longer increase the weight of progress；But regrettably: China at least 90% POAG and at least 50% PACG not by
Find or patient oneself does not knows suffer from glaucoma.During until patient actively goes to see a doctor, majority be have in, late period regard god
Through the glaucoma patient of infringement, the infringement of optic nerve form and visual field damage ratio are more serious, and the even visual field will be lost completely
And it is blind.
Glaucoma screening mode includes opportunistic examination and Mass screening two aspect:
Opportunistic examination refers to some because health examination or other problems carry out ophthalmology when going to a doctor, and ophthalmologists enters consciously
Necessary inspection in terms of row glaucoma, surveys including examination with slitlamp microscope peripheral chamber depth, optical fundus digital photographing, intraocular pressure
Amount, carries out gonioscope etc. if desired and checks, it is a part for clinical ophthalmology work.It is applicable to Ophthalmologic apparatus and medical worker
The tertiary hospitals that quality is higher.
Mass screening is the examination carried out in ordinary populace or special group, big towards crowd's radix, participates in generaI investigation
Equipment and peopleware irregular have different, but be most important in glaucoma-controlling be also widest diagnosis and treatment First Line.
Current existing diagnosis of glaucoma means have the most several:
(1) tonometry: use detection of eyeball tension as the method for examination glaucoma, it is difficult to determine normal intraocular tension threshold line, by eye
Pressure > 21 mmHg are as boundary, and its sensitivity is 47%, and specificity is 92%(Baltimore Eye Survey, 1991), i.e. have
The early-stage glaucoma patient exceeding half is the highest at random detection intraocular pressure in the daytime.Therefore, modern diagnosis of glaucoma introduce because of
People and different target intraocular pressure concept and personalized therapy program.
(2) standardization automated threshold: although checking the goldstandard being to assess glaucoma functional lesion, but it belongs to subjective inspection
Looking into, rely on experimenter and cooperate, the review time is longer, and (current SITA-Fast, TOP checks that program is the shortest, every inspection
Look at least with the time of a few minutes), additionally basic unit's popularity rate of visual field instrument is relatively low, and volume is the most not readily portable, and these all limit
Make it and carry out extensive basic unit Mass screening.
(3) frequency multiplication perimetry (FDP) is used: although review time shorter (full threshold modal every 5 minutes；Above threshold it is worth
Examination pattern be less than 90 seconds), but its centering, severe glaucoma detection more sensitive.
(4) directly ophthalmofundoscope: evaluate optic nerve and there is quick, economic advantage, but its sensitivity and specificity is relatively low, is subject to
Medical worker's quality influence error is relatively big, lacks objective record, is unfavorable for the consultation of doctors and Database.
(5) optic nerve quantitative measurement instrument such as HRT, GDX, OCT etc. check: the time is longer, and inspection apparatus price is high, basic unit
Popularity rate is the lowest, and inspection personnel's specialty requires height, and result is explained needs specific Professional knowledge etc., to most Mass screenings not
Practical.
In order to improve the effect/cost ratio of glaucoma Mass screening, the problem that should be noted three aspects.One is should be for mesh
Mark Mass screening, two is associating many sick kinds examination simultaneously in addition to glaucoma, and three is to set up ophthalmoscopic image data base, forms expansion
GeneraI investigation diagnosis and therapy system.Because so-called target group refers mainly to glaucoma high-risk group, including more than 40 years old person, diabetics
(it suffers from glaucoma probability the 2.12 of the person that is non-diabetic research display), the high myopic eye of myopia especially more than 6D are suffered from
Person's (have research display its to suffer from glaucoma probability be emmetropic 7.56 times), glaucoma family history person (have research display POAG mono-
Level relatives' risk is 10%).Multiple disease United screening refers in addition to glaucoma, the person that cataract can be needed operative treatment,
Multiple blinding oculopathy such as diabetic retinopathy, age-related macular degeneration, the change of pathologic myopia optical fundus and complete
The retinal complication of body disease (such as hypertension, hyperlipemia, apoplexy etc.) slowly carries out examination simultaneously.
Medical research confirms, optic nerve progressive injury is the pathologic basis of glaucoma blinding.More meaningful, green light
Eye optic nerve lesion is different from other kinds of optic nerve lesion, it with optic disc dish along lose and corresponding site with dish along being connected
Retinal nerve fibre layer defect be characterized, cup disc ratio value produces change therewith, and this characteristic pathological is changed to glaucoma
Examination provides foundation and facility.By fundus imaging being measured the cup disc ratio value (C/D ratio) of papilla of optic nerve, it is for early
Phase finds one of method of glaucoma, the most day by day recognized by ophthalmologists.But, use easy measuring appliance to measure cup dish
More complicated than operation, the specialty of inspection personnel is required and clinical experience requires higher, use the measuring instrument then price of complexity
High, both of which cannot be promoted at basic medical unit effectively.China's medical research confirms, glaucoma in current Chinese population
Prevalence is high, and diagnosis and treatment efficiency is low, this make to research and develop a kind of cost performance high (i.e. glaucoma patient detection number/generaI investigation input cost),
Crowd's glaucoma screening system that high popularity rate, expansibility are strong has extremely important meaning.
The vast basic hospital deficienter for Ophthalmologic apparatus and professional and amateur ophthalmology prophylactico-therapeutic institution, it is provided that one
Kind cost performance popularity rate high, high, the glaucoma methods for screening that expansibility is strong, carry out low-cost high-efficiency to fundus photograph
Examination, improve glaucoma patient positive rate, it is achieved in crowd the morning of glaucoma find, early diagnosis, early treatment, and progressively
Improve big data expansion development system under three cascade network systems the most meaningful.
Summary of the invention
It is an object of the invention to: a kind of cost performance popularity rate high, high, the glaucoma early screening that expansibility is strong are provided
Method.
The purpose of the present invention is realized by following proposal: the methods for screening of a kind of glaucoma, utilizes computer to eye
End photo carries out automatical analysis, calculates the cup disc ratio C/D ratio of optic cup and optic disc the most in real time, wherein, by cup disc ratio with
The glaucoma value-at-risk contrast set judges whether the probability of glaucoma, is realized by following steps:
Step 1, obtains the original fundus photograph of RGB color pattern by fundus imaging equipment, inputs computer；
Step 2, includes hsv color patten transformation, and extracts the pre-place of luminance parameter V therein original fundus photograph
Reason, it is thus achieved that optical fundus brightness histogram；
Step 3, carries out binaryzation and forms black and white image optical fundus brightness histogram, uses morphological image method to extract
UNICOM's administrative division map, demarcates optic disc region by optic disc region screening method, it is thus achieved that optic disc boundary graph, intercepts in original fundus photograph
Optic disc region photo；
Step 4, includes RGB gradation conversion method and the pretreatment of image blurring process to optic disc region photo, it is thus achieved that optic disc
Area grayscale fuzzy graph；
Step 5, calculates the intensity histogram diagram data of optic disc area grayscale fuzzy graph, and optic disc area grayscale fuzzy graph is used OTSU
Algorithm obtains cup boundary figure；
Step 6, the optic disc border obtained according to above-mentioned steps 3 and step 5 and cup boundary, use minimum circumscribed circle algorithm respectively
Calculate the minimum circumscribed circle of optic disc and optic cup, and calculate cup disc ratio by the minimum circumscribed circle diameter of optic disc and optic cup；
Step 7, by the glaucoma value-at-risk contrast of cup disc ratio with setting, it is thus achieved that comparing result, computer export cup disc ratio numerical value
Comparing result with the glaucoma value-at-risk set.
On the basis of such scheme, described binaryzation utilizes luminance threshold that optical fundus brightness histogram is divided into black and white two
The image of color, its computing formula is:
T=round(Imax-(Imax-Imin)/2)*K
Wherein, T is the luminance threshold for image binaryzation, is white more than T, is black less than T, and K is fundus imaging equipment
Luminance threshold, the value of K is relevant with the hardware parameter of fundus imaging equipment, and the experiment that can pass through obtains, and round () is
The canonical function of MATLAB software, Imax is the maximum of brightness in the brightness/gray scale image of optical fundus, and Imin is optical fundus brightness/gray scale
The minima of brightness in image.
RGB gradation conversion be according to shooting fundus photograph used by fundus imaging equipment to redgreenblue sensitivity
Export ratio, applies below equation to change:
Gray = R*Sr+G*Sg+B*Sb
Thus obtain the gray scale picture of correspondence.
Wherein, Sr, Sg and Sb are respectively the camera output photo sensitivity ratio to three kinds of colors.
Owing to every kind of model camera or same portion camera shooting mode are different, the sensitivity output to redgreenblue is all deposited
At fine difference.Therefore, Sr, Sg and Sb typically, before equipment is applied, by the shooting of the fundus photograph to multiple cases, and try
The ratio (Sr+Sg+Sb=1) of a kind of best results is drawn after calculation.The most various acquisition parameters of fixed camera, with
Time fix Sr, Sg and Sb export ratio.
On the basis of such scheme, described optic disc region screening method includes area exclusive method and circle degree of approximation sequence
Method, area exclusive method is area or the pixel quantity calculating UNICOM region, and area or pixel quantity are less than human eye optic disc
The region of the 68% of average area or the 68% of human eye optic disc photo mean pixel point quantity is got rid of, and then uses circle degree of approximation sequence
Method is to calculate UNICOM's region maximum containing circular diameter to determine with minimum circumscribed circle diameter ratio value, and is arranged from high to low by ratio
Sequence, the UNICOM region that ratio is the highest is optic disc region.
On the basis of such scheme, described image blurring process is defocusing blurring, Gaussian Blur or average blur.Make
Reason with image blurring process: due in the figure of optical fundus in addition to optic cup, optic disc, there is also the image of big blood vessel, and blood vessel
Border is sharper keen with what the border of optic disc presented on the figure of optical fundus relative to optic cup.If directly using OTSU method to ask for optic cup
Border, it is easy to affected by these optical fundus blood vessels and produce so-called boundary leaking effect.Accordingly, it would be desirable to carry out image mould
Paste processes, and weakens the boundary effect of optical fundus blood vessel.
3 kinds of conventional image blurring processing methods of MATLAB are respectively defocusing blurring, Gaussian Blur or average blur:
r=10；(filter radius parameter r)
Defocusing blurring:
PSF1=fspecial('disk',r)；(border circular areas mean filter)
I1=imfilter (Im1, PSF1, ' symmetric', ' conv') (original image Im1 exports image I1)
Gaussian Blur:
PSF2=fspecial('gaussian',r)；(Gassian low-pass filter)
I2=imfilter (Im2, PSF2, ' symmetric', ' conv') (original image Im2 exports image I2)
Average blur:
PSF3=fspecial('average',r)；(mean filter)
I3=imfilter (Im3, PSF3, ' symmetric', ' conv') (original image Im3 exports image I3)
Through lot of experiments, above-mentioned three kinds of fuzzy algorithmic approaches are compared, defocusing blurring relative to other two kinds of fuzzy algorithmic approaches, its
The precision of result is higher and fault-tolerance more preferable, is suitable for the various Parameters variation of fundus photograph in actual application, as optical fundus is shone
The parameters such as the color saturation of sheet, brightness.
Defocusing blurring process after feature be that image starts from center radially to obscure.This algorithm can greatly weaken
The boundary effect parallel with radiation direction, and the border vertical with radiation direction is weakened less.And optical fundus blood vessel is substantially
From eye center radially to extend near the eyes, therefore after defocusing blurring processes, the vessel borders on the figure of optical fundus is just by pole
Big reduction.
The parameter of fuzzy algorithmic approach: the result through experiment with OCT optical fundus tomoscan is calibrated and determined, such as fundus photograph
Color saturation, brightness and filter radius parameter r etc..
On the basis of such scheme, described intensity histogram diagram data is the canonical function by MATLAB software
Data after statistics are shown as coordinate diagram by the data in imhist () statistical picture, and its abscissa represents the gray level of pixel
Not, vertical coordinate is the number of pixel, and this coordinate diagram is grey level histogram.Owing to OTSU algorithm is that the gray scale to image is grading
Row cluster, so before performing OTSU algorithm, needing to calculate the intensity histogram diagram data of this image.
On the basis of such scheme, described OTSU algorithm utilizes gray threshold to be divided into by optic disc area grayscale fuzzy graph
Prospect, two images of background, its formula is:
G=w0* (u0-u) * (u0-u)+w1* (u1-u) * (u1-u), or
g=w0*w1*(u0-u1)*(u0-u1)
Wherein, if the gray threshold that t is prospect and background, gray value less than gray threshold t for prospect, gray value is more than gray scale
Threshold value t for background, foreground pixel is counted and accounted for image scaled is w0, and average gray is u0, and background pixel is counted and accounted for image scaled
For w1, average gray be u1, u be image grand mean gray scale, u=w0*u0+w1*u1, g are the variance of foreground and background image,
The method using traversal t obtains the gray threshold t making variance g maximum, is required gray threshold.
OTSU method (maximum variance between clusters, the biggest law) uses the thought of cluster, image
Grey is divided into 2 parts by gray level so that the grey value difference between two parts is maximum, the gray scale between each part
Difference is minimum, finds a suitable grey level by the calculating of variance and divides.So can be binaryzation when
Use OTSU algorithm to carry out automatic selected threshold and carry out binaryzation.OTSU algorithm be considered as image segmentation in threshold value choose optimal
Algorithm, calculates simple, is not affected by brightness of image and contrast.Therefore, the segmentation making inter-class variance maximum means wrong point
Probability is minimum.Corresponding to the segmentation of optic cup optic disc, OTSU method can obtain the optimal computed result under prosthetic aided case.
On the basis of such scheme, in step 6, described minimum circumscribed circle algorithm is following four step:
Step 6.1, appoints in the pixel in region and takes A, B, C at 3.；
Step 6.2, makees a smallest circle comprising A, B, C 3, and circumference is by 3 or only by wherein 2 point, only by it
In 2 time need to comprise thirdly；
Step 6.3, finds out distance step 6.2 and is built the D point that the round heart is farthest, if D point is being justified in the pixel in region
In or circumference on, then this circle is required circle, and algorithm terminates, and otherwise performs next step；
Step 6.4, selects three points in A, B, C, D, and it is minimum for making the circle comprising these 4 points generated by them, and these are three years old
Point becomes new A, B, C, if the circumference of the circle generated is only by 2 points in A, B, C, D, then 2 on circumference take into new A
And B, appoint from another 2 and take a little as new C, return and perform step 6.2.
On the basis of such scheme, described original fundus photograph, cup disc ratio numerical value and the glaucoma risk with setting
The comparing result of value can be uploaded to higher level data base by LAN or the Internet, and can ultimately form big data analysis sample
This.Large-scale glaucoma is generally investigated and sets up the image database of the Internet, it is simple to remote medical consultation with specialists, forms three grades of layers even, different
Medical institutions, the ophthalmologists of different location can be checked from network by login password, and diagnosis and treatment suggestion be exchanged with patient,
It is beneficial to examining, compares with baseline screening results, thus monitor disease progression.
By passing through the different medical unit (community in urban areas or villages and towns in rural areas hospital etc.) of training or joining mechanism (such as healthy body
Inspection center, non-ophthalmology section hospital etc.) medical related personnel publicized by the means such as network, broadcast, TV in locality, tissue
Examination, in particular for target group, examination of visual acuity, carries out exempting from mydriasis optical fundus digital photographing, and every carries out including optic disc and Huang
Macular area, an interior fundus image, the fundus image person of not knowing, need to take anterior ocular segment picture with fundus camera, to understand anterior ocular segment situation
And ocular media opacity especially cataract degree.Image carries out computing through the method for the present invention, quickly obtains cup disc ratio, i.e.
Preliminary examination and the diagnosis of glaucoma can be carried out in real time.Parameter to tentative diagnosis case and suspected case uploads to interconnection simultaneously
Grid database, by Grade III Class A hospital's real-time diagosis of glaucoma specialist, uploads diagosis report, basic unit's cooperation unit in real time
Medical worker and experimenter can check picture and diagosis result on the internet.Diagosis doctor is according to cup disc ratio and retina
The morphological change of nerve fibre layer such as top or disk under edge narrows, eyes optic cup is asymmetric, limitation retina neural fine
Dimension break damages, dish combines medical history make the diagnosis of glaucoma and suspected glaucoma along hemorrhage etc., and further treats suggestion.
Advantages of the present invention: produce cup disc ratio with optic nerve Progressive symmetric erythrokeratodermia damage in glaucoma pathological changes and be changed to basis, in conjunction with
Fundus photography and the optic cup researched and developed voluntarily, optic disc boundary detection method, it is achieved full-automatic fundus photograph diagosis under area of computer aided
Analyze.Optic cup under the color picture system of optical fundus, optic disc boundary detection method, improve the early detective rate of glaucoma, evade simultaneously
Detection equipment costly, equipment needed thereby low cost, use simple, operator's specialty is required low, by optical fundus
Photo carries out the examination of low-cost high-efficiency, it is achieved in crowd, the morning of glaucoma finds, morning diagnoses, early treatment；Due to image mould
Paste process, OTSU algorithm and the introducing of Computing, compared with the method in the past utilizing merely fundus photography diagosis to diagnose, greatly
Improve greatly accuracy and real-time that result judges, it is simple to the foundation of data base and further follow-up investigation, substantially increase
The effect of Mass screening/cost ratio, and can gradual perfection three cascade by the way of step-sizing uploads higher level data base step by step
Big data expansion development system under net system.
Accompanying drawing explanation
Fig. 1 is glaucoma early intervention and the difference intervening final result late period；
Fig. 2 is the glaucoma screening method flow diagram of the present invention；
Fig. 3 is the original fundus photograph of left eye of the present invention；
Fig. 4 is the original fundus photograph of right eye of the present invention；
Fig. 5 is the left eye optical fundus brightness histogram of the present invention；
Fig. 6 is the right eye optical fundus brightness histogram of the present invention；
Fig. 7 is the left eye optical fundus brightness histogram after the binaryzation of the present invention；
Fig. 8 is the right eye optical fundus brightness histogram after the binaryzation of the present invention；
Fig. 9 is the left eye optic disc boundary graph of the present invention；
Figure 10 is the right eye optic disc boundary graph of the present invention；
Figure 11 is the left eye optic disc region photo of the present invention；
Figure 12 is the right eye optic disc region photo of the present invention；
Figure 13 is the left eye optic disc area grayscale fuzzy graph of the present invention；
Figure 14 is the right eye optic disc area grayscale fuzzy graph of the present invention；
Figure 15 is the left eye grey level histogram of the present invention；
Figure 16 is the right eye grey level histogram of the present invention；
Figure 17 is the left eye cup boundary figure of the present invention；
Figure 18 is the right eye cup boundary figure of the present invention.
Detailed description of the invention
Elaborating embodiments of the invention below in conjunction with the accompanying drawings, the fundus imaging equipment that the present embodiment is used is
Canon CR-DGi(exempts from mydriasis fundus camera), used software is MATLAB, such as the glaucoma screening method stream of Fig. 2 present invention
Shown in journey figure, pass through following steps.
Step 1, by the original eye of images of left and right eyes of fundus imaging equipment acquisition RGB color pattern as shown in Figure 3 and Figure 4
End photo, inputs computer；
Step 2, carries out hsv color patten transformation, and extracts luminance parameter V therein original fundus photograph, it is thus achieved that such as Fig. 5 and
Images of left and right eyes optical fundus brightness histogram shown in Fig. 6；
Step 3, after carrying out the black and white binaryzation that binaryzation is formed as shown in Figure 7 and Figure 8 to optical fundus brightness histogram
Images of left and right eyes optical fundus brightness histogram, uses morphological image method to extract UNICOM's administrative division map, by area exclusive method and circle approximation
Degree ranking method demarcates optic disc region, by the pixel quantity region row less than the 68% of human eye optic disc photo mean pixel point quantity
Removing, then using circle degree of approximation ranking method is to calculate UNICOM's region maximum containing circular diameter to come with minimum circumscribed circle diameter ratio value
Determining, and sorted from high to low by ratio, the UNICOM region that ratio is the highest is optic disc region, it is thus achieved that as shown in Figure 9 and Figure 10
Images of left and right eyes optic disc boundary graph, intercept the images of left and right eyes optic disc region photo in original fundus photograph as is illustrated by figs. 11 and 12,
The computing formula of binaryzation is:
T=round(Imax-(Imax-Imin)/2)*K
Wherein, T is the luminance threshold for image binaryzation, is white more than T, is black less than T, and K is fundus imaging equipment
Luminance threshold, in the present embodiment, K takes 0.92, and round () is the canonical function of MATLAB software, and Imax is optical fundus brightness/gray scale
The maximum of brightness in image, Imin is the minima of brightness in the brightness/gray scale image of optical fundus；
Step 4, carries out RGB gradation conversion and defocusing blurring to optic disc region photo, it is thus achieved that as shown in Figure 13 and Figure 14 left and right
Eye optic disc area grayscale fuzzy graph；
Data after statistics, by the data in canonical function the imhist () statistical picture of MATLAB software, are shown by step 5
For coordinate diagram, its abscissa represents the grey level of pixel, and vertical coordinate is the number of pixel, as shown in Figure 15 and Figure 16
Images of left and right eyes grey level histogram, uses OTSU algorithm to obtain such as figure optic disc area grayscale fuzzy graph according to intensity histogram diagram data
Images of left and right eyes cup boundary figure shown in 17 and Figure 18；
Step 6, the optic disc border obtained according to above-mentioned steps 3 and step 5 and cup boundary, use minimum circumscribed circle algorithm respectively
Calculate the minimum circumscribed circle of optic disc and optic cup, and calculate cup disc ratio by the minimum circumscribed circle diameter of optic disc and optic cup；
Step 7, contrasts the glaucoma value-at-risk 0.3 of cup disc ratio with setting, and computer export comparing result is less than when cup disc ratio
When 0.3, prompting is normal, and what when cup disc ratio is more than 0.3, glaucoma was suffered from prompting may be higher, needs detection further, and will contrast
Result is uploaded to higher level data base by LAN or the Internet, ultimately forms big data analysis sample.
Claims (8)
1. a methods for screening for glaucoma, utilizes computer that fundus photograph is carried out automatical analysis, the most in real time meter
Calculate the cup disc ratio C/D ratio of optic cup and optic disc, it is characterised in that the glaucoma value-at-risk contrast of cup disc ratio with setting judged
Whether there is the probability of glaucoma, realized by following steps:
Step 1, obtains the original fundus photograph of RGB color pattern by fundus imaging equipment, inputs computer；
Step 2, carries out hsv color patten transformation to original fundus photograph, and extracts the pretreatment of luminance parameter V therein, obtains
Obtain optical fundus brightness histogram；
Step 3, carries out binaryzation and forms black and white image optical fundus brightness histogram, uses morphological image method to extract
UNICOM's administrative division map, demarcates optic disc region by optic disc region screening method, it is thus achieved that optic disc boundary graph, intercepts in original fundus photograph
Optic disc region photo；
Step 4, includes RGB gradation conversion method and the pretreatment of image blurring process to optic disc region photo, it is thus achieved that optic disc
Area grayscale fuzzy graph；
Step 5, calculates the intensity histogram diagram data of optic disc area grayscale fuzzy graph, and optic disc area grayscale fuzzy graph is used OTSU
Algorithm obtains cup boundary figure；
Step 6, the optic disc border obtained according to above-mentioned steps 3 and step 5 and cup boundary, use minimum circumscribed circle algorithm respectively
Calculate the minimum circumscribed circle of optic disc and optic cup, and calculate cup disc ratio by the minimum circumscribed circle diameter of optic disc and optic cup；
Step 7, by the glaucoma value-at-risk contrast of cup disc ratio with setting, it is thus achieved that comparing result, computer export cup disc ratio numerical value
Comparing result with the glaucoma value-at-risk set.
The methods for screening of a kind of glaucoma the most according to claim 1, it is characterised in that: described binaryzation utilizes
Optical fundus brightness histogram is divided into black and white image by luminance threshold, and its computing formula is:
T=round(Imax-(Imax-Imin)/2)*K
Wherein, T is the luminance threshold for image binaryzation, is white more than T, is black less than T, and K is fundus imaging equipment
Luminance threshold, round () is the canonical function of MATLAB software, and Imax is the maximum of brightness in the brightness/gray scale image of optical fundus
Value, Imin is the minima of brightness in the brightness/gray scale image of optical fundus.
The methods for screening of a kind of glaucoma the most according to claim 1, it is characterised in that: described optic disc region
Screening method includes area exclusive method and circle degree of approximation ranking method, and area exclusive method is to calculate area or the pixel number in UNICOM region
Amount, is less than 68% or human eye optic disc photo mean pixel point quantity of human eye optic disc average area by area or pixel quantity
The region of 68% is got rid of, and then using circle degree of approximation ranking method is to calculate UNICOM's region maximum to contain circular diameter and minimum circumscribed circle
Diameter ratio value determines, and is sorted from high to low by ratio, and the UNICOM region that ratio is the highest is optic disc region.
The methods for screening of a kind of glaucoma the most according to claim 1, it is characterised in that: described image blurring place
Reason is defocusing blurring, Gaussian Blur or average blur.
The methods for screening of a kind of glaucoma the most according to claim 1, it is characterised in that: described grey level histogram
Data are by the data in canonical function the imhist () statistical picture of MATLAB software, are shown as sitting by the data after statistics
Marking on a map, its abscissa represents the grey level of pixel, and vertical coordinate is the number of pixel, and this coordinate diagram is grey level histogram.
A kind of methods for screening of glaucoma, it is characterised in that: described OTSU calculates
Method utilizes gray threshold that optic disc area grayscale fuzzy graph is divided into prospect, two images of background, and its formula is:
G=w0* (u0-u) * (u0-u)+w1* (u1-u) * (u1-u), or
g=w0*w1*(u0-u1)*(u0-u1)
Wherein, if the gray threshold that t is prospect and background, gray value less than gray threshold t for prospect, gray value is more than gray scale
Threshold value t for background, foreground pixel is counted and accounted for image scaled is w0, and average gray is u0, and background pixel is counted and accounted for image scaled
For w1, average gray be u1, u be image grand mean gray scale, u=w0*u0+w1*u1, g are the variance of foreground and background image,
The method using traversal t obtains the gray threshold t making variance g maximum, is required gray threshold.
The methods for screening of a kind of glaucoma the most according to claim 1, it is characterised in that: in step 6, described
Minimum circumscribed circle algorithm is following four step:
Step 6.1, appoints in the pixel in region and takes A, B, C at 3.；
Step 6.2, makees a smallest circle comprising A, B, C 3, and circumference is by 3 or only by wherein 2 point, only by it
In 2 time need to comprise thirdly；
Step 6.3, finds out distance step 6.2 and is built the D point that the round heart is farthest, if D point is being justified in the pixel in region
In or circumference on, then this circle is required circle, and algorithm terminates, and otherwise performs next step；
Step 6.4, selects three points in A, B, C, D, and it is minimum for making the circle comprising these 4 points generated by them, and these are three years old
Point becomes new A, B, C, if the circumference of the circle generated is only by 2 points in A, B, C, D, then 2 on circumference take into new A
And B, appoint from another 2 and take a little as new C, return and perform step 6.2.
The methods for screening of a kind of glaucoma the most according to claim 1, it is characterised in that: described original optical fundus is shone
Sheet, cup disc ratio numerical value and the comparing result with the glaucoma value-at-risk set can be uploaded to higher level by LAN or the Internet
Data base, and big data analysis sample can be ultimately formed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610689336.2A CN106214120A (en) | 2016-08-19 | 2016-08-19 | A kind of methods for screening of glaucoma |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610689336.2A CN106214120A (en) | 2016-08-19 | 2016-08-19 | A kind of methods for screening of glaucoma |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106214120A true CN106214120A (en) | 2016-12-14 |
Family
ID=57553583
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610689336.2A Pending CN106214120A (en) | 2016-08-19 | 2016-08-19 | A kind of methods for screening of glaucoma |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106214120A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109684981A (en) * | 2018-12-19 | 2019-04-26 | 上海鹰瞳医疗科技有限公司 | Glaucoma image-recognizing method, equipment and screening system |
| CN109697716A (en) * | 2018-12-19 | 2019-04-30 | 上海鹰瞳医疗科技有限公司 | Glaucoma image-recognizing method, equipment and screening system |
| CN109712693A (en) * | 2018-12-14 | 2019-05-03 | 贵州大学 | A kind of intelligence pathological diagnosis method and system |
| WO2019218482A1 (en) * | 2018-05-14 | 2019-11-21 | 平安科技（深圳）有限公司 | Big data-based population screening method and apparatus, terminal device and readable storage medium |
| CN111080630A (en) * | 2019-12-20 | 2020-04-28 | 腾讯医疗健康(深圳)有限公司 | Fundus image detection apparatus, method, device, and storage medium |
| CN112765563A (en) * | 2020-10-24 | 2021-05-07 | 昆明医科大学 | Method for assessing glaucoma progression based on retinal ganglion cell estimates |
| CN113780477A (en) * | 2021-10-11 | 2021-12-10 | 深圳硅基智能科技有限公司 | Method and device for measuring fundus image based on deep learning of tight frame mark |
| US11295178B2 (en) | 2017-11-01 | 2022-04-05 | Tencent Technology (Shenzhen) Company Limited | Image classification method, server, and computer-readable storage medium |
-
2016
- 2016-08-19 CN CN201610689336.2A patent/CN106214120A/en active Pending
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11295178B2 (en) | 2017-11-01 | 2022-04-05 | Tencent Technology (Shenzhen) Company Limited | Image classification method, server, and computer-readable storage medium |
| WO2019218482A1 (en) * | 2018-05-14 | 2019-11-21 | 平安科技（深圳）有限公司 | Big data-based population screening method and apparatus, terminal device and readable storage medium |
| CN109712693A (en) * | 2018-12-14 | 2019-05-03 | 贵州大学 | A kind of intelligence pathological diagnosis method and system |
| WO2020125319A1 (en) * | 2018-12-19 | 2020-06-25 | 上海鹰瞳医疗科技有限公司 | Glaucoma image recognition method and device and screening system |
| WO2020125318A1 (en) * | 2018-12-19 | 2020-06-25 | 上海鹰瞳医疗科技有限公司 | Glaucoma image recognition method and device and diagnosis system |
| CN109684981A (en) * | 2018-12-19 | 2019-04-26 | 上海鹰瞳医疗科技有限公司 | Glaucoma image-recognizing method, equipment and screening system |
| CN109697716B (en) * | 2018-12-19 | 2021-04-02 | 上海鹰瞳医疗科技有限公司 | Identification method and equipment of cyan eye image and screening system |
| CN109697716A (en) * | 2018-12-19 | 2019-04-30 | 上海鹰瞳医疗科技有限公司 | Glaucoma image-recognizing method, equipment and screening system |
| CN111080630A (en) * | 2019-12-20 | 2020-04-28 | 腾讯医疗健康(深圳)有限公司 | Fundus image detection apparatus, method, device, and storage medium |
| CN111080630B (en) * | 2019-12-20 | 2024-03-08 | 腾讯医疗健康(深圳)有限公司 | Fundus image detection device, fundus image detection method, fundus image detection device, and fundus image storage medium |
| CN112765563A (en) * | 2020-10-24 | 2021-05-07 | 昆明医科大学 | Method for assessing glaucoma progression based on retinal ganglion cell estimates |
| CN113780477A (en) * | 2021-10-11 | 2021-12-10 | 深圳硅基智能科技有限公司 | Method and device for measuring fundus image based on deep learning of tight frame mark |
| CN113780477B (en) * | 2021-10-11 | 2022-07-22 | 深圳硅基智能科技有限公司 | Method and device for measuring fundus image based on deep learning of tight frame mark |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106214120A (en) | A kind of methods for screening of glaucoma | |
| Cao et al. | Hierarchical method for cataract grading based on retinal images using improved Haar wavelet | |
| Zhang et al. | A survey on computer aided diagnosis for ocular diseases | |
| Gao et al. | Automatic feature learning to grade nuclear cataracts based on deep learning | |
| Markowitz et al. | Microperimetry and clinical practice: an evidence-based review | |
| Patton et al. | Retinal image analysis: concepts, applications and potential | |
| Nayak et al. | Automated identification of diabetic retinopathy stages using digital fundus images | |
| Xiong et al. | An approach to evaluate blurriness in retinal images with vitreous opacity for cataract diagnosis | |
| US20120150029A1 (en) | System and Method for Detection and Monitoring of Ocular Diseases and Disorders using Optical Coherence Tomography | |
| WO2020042500A1 (en) | Health big data service method and system based on remote fundus screening | |
| WO2012136079A1 (en) | Method and device for retinal image analysis | |
| CN111785363A (en) | AI-guidance-based chronic disease auxiliary diagnosis system | |
| Shao et al. | Quantitative assessment of fundus tessellated density and associated factors in fundus images using artificial intelligence | |
| WO2010131944A2 (en) | Apparatus for monitoring and grading diabetic retinopathy | |
| Adalarasan et al. | Automatic detection of blood vessels in digital retinal images using soft computing technique | |
| Naveed et al. | Clinical and technical perspective of glaucoma detection using OCT and fundus images: a review | |
| Wang et al. | Accurate disease detection quantification of iris based retinal images using random implication image classifier technique | |
| CN111402184B (en) | Method and system for realizing remote fundus screening and health service | |
| Giancardo | Automated fundus images analysis techniques to screen retinal diseases in diabetic patients | |
| de Moura et al. | Automatic characterization of the serous retinal detachment associated with the subretinal fluid presence in optical coherence tomography images | |
| Consejo et al. | Detection of subclinical keratoconus with a validated alternative method to corneal densitometry | |
| Frampton et al. | Accuracy of fundus autofluorescence imaging for the diagnosis and monitoring of retinal conditions: a systematic review | |
| Ajaz et al. | Association between Optical Coherence Tomography and Fluorescein Angiography based retinal features in the diagnosis of Macular Edema | |
| Tugashetti et al. | Design of algorithms for diagnosis of primary glaucoma through estimation of CDR in different types of fundus images using IP techniques | |
| WO2019157113A1 (en) | Segmentation-based corneal mapping |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20161214 |
|
| WD01 | Invention patent application deemed withdrawn after publication |