CN109658395A - Optic disk method for tracing and system and eyeground acquisition device - Google Patents
Optic disk method for tracing and system and eyeground acquisition device Download PDFInfo
- Publication number
- CN109658395A CN109658395A CN201811485787.XA CN201811485787A CN109658395A CN 109658395 A CN109658395 A CN 109658395A CN 201811485787 A CN201811485787 A CN 201811485787A CN 109658395 A CN109658395 A CN 109658395A
- Authority
- CN
- China
- Prior art keywords
- optic disk
- centroid position
- fundus image
- eye fundus
- noise
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 210000003733 optic disk Anatomy 0.000 title claims abstract description 257
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000004364 calculation method Methods 0.000 claims description 15
- 238000000605 extraction Methods 0.000 claims description 10
- 230000000052 comparative effect Effects 0.000 claims description 6
- 210000003128 head Anatomy 0.000 claims description 6
- 238000003333 near-infrared imaging Methods 0.000 abstract description 2
- 230000004807 localization Effects 0.000 abstract 1
- 238000003384 imaging method Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 210000004204 blood vessel Anatomy 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 210000002189 macula lutea Anatomy 0.000 description 1
- 231100000915 pathological change Toxicity 0.000 description 1
- 230000036285 pathological change Effects 0.000 description 1
- 210000001525 retina Anatomy 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- 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
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10048—Infrared image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30041—Eye; Retina; Ophthalmic
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Eye Examination Apparatus (AREA)
Abstract
The invention discloses a kind of optic disk method for tracing, the optic disk centroid position including calculating each frame eye fundus image in near-infrared fundus video stream;Initial preceding N frame eye fundus image and optic disk centroid position are extracted, and determines current judgement queue initial frame;Calculate the distance between the optic disk centroid position of the variance of the optic disk centroid position of present frame eye fundus image and the optic disk centroid position of previous valid frame eye fundus image and the optic disk centroid position of present frame eye fundus image and benchmark optic disk;Determine whether present frame eye fundus image is noise according to variance and reference distance;When determining as noise, present frame eye fundus image is rejected;It determines when not being noise, caches present frame eye fundus image;Continue whether next frame eye fundus image is the judgement of noise, until eliminating earliest frame when the current length for determining queue reaches threshold value according to the principle of first in first out and introducing new frame.Not the problem of not being suitable for dynamic, near infrared imaging eyeground figure which solve traditional optic disk localization method.
Description
Technical field
The present invention relates to technical field of medical image processing, more particularly to a kind of optic disk method for tracing and system and eyeground
Acquisition device.
Background technique
Key element one of of the optic disk as fundus imaging figure is cured with blood vessel, macula lutea, retina collectively as eyeground pathological changes
Learn the key element that diagnosis relies on.The currently employed method and apparatus for carrying out image recognition for optic disk, be mostly for
Static, the eyeground figure after visual light imaging is analyzed, and optic disk, complementary medicine diagnosis are extracted.And for dynamic, near-infrared
The eyeground figure of imaging is then not suitable for.
Summary of the invention
Based on this, it is necessary to for traditional optic disk recognition positioning method and device be not suitable for dynamic, near-infrared at
The problem of eyeground figure of picture, provides a kind of optic disk method for tracing and system and eyeground acquisition device.
Based on above-mentioned purpose, a kind of optic disk method for tracing provided by the invention, which comprises the steps of:
Collected near-infrared fundus video stream is obtained, each frame eye fundus image in the near-infrared fundus video stream is calculated
Optic disk centroid position;
Preceding N frame eye fundus image and its optic disk centroid position initial in the near-infrared video stream are extracted, and according to extraction
The optic disk centroid position by determining the current initial frame for determining queue in the preceding N frame eye fundus image;Wherein, described current
The regular length threshold value for determining queue is n;
For the present frame eye fundus image for currently determining newly to capture in queue, the view of the present frame eye fundus image is calculated
The variance of the optic disk centroid position of disk centroid position and previous valid frame eye fundus image and the view of the present frame eye fundus image
The distance between the optic disk centroid position of disk centroid position and benchmark optic disk;Wherein, the optic disk matter of the present frame eye fundus image
Distance on the basis of the distance between optic disk centroid position of heart position and benchmark optic disk;
The optic disk mass center of the present frame eye fundus image is determined according to the variance and the reference distance that are calculated
Whether position is noise;And when the optic disk centroid position for determining the present frame eye fundus image is noise, work as described in rejecting
Previous frame eye fundus image;When the optic disk centroid position for determining the present frame eye fundus image is not noise, the present frame is cached
Eye fundus image and its optic disk centroid position;
Continue to capture next frame eye fundus image in current judgement queue and is determined, until the current judgement queue
After length reaches threshold value n, then for each frame eye fundus image of subsequent captured, into while the current judgement queue, root
According to the principle of first in first out, the frame of queue head is eliminated.
It is described in one of the embodiments, to extract preceding N frame eye fundus image and institute initial in the near-infrared video stream
Optic disk centroid position is stated, and according to the optic disk centroid position of extraction by determining current judgement in the preceding N frame eye fundus image
The initial frame of queue, includes the following steps:
By N frame eye fundus image and its corresponding N number of optic disk mass center position before being extracted in the near-infrared fundus video stream
It sets;
The abscissa to N number of optic disk centroid position and ordinate carry out mean value calculation respectively, obtain average mass center
Position;
It calculates in N number of optic disk centroid position between each described optic disk centroid position and the average centroid position
Distance, obtain N number of average distance;
Size sequence is carried out to N number of average distance, and rejects being averaged for predetermined number according to sequence from big to small
After distance, N-Q reasonable average distances are obtained;Wherein, Q is predetermined number;
The maximum average range in the reasonable average distance is chosen as boundary value, and on the basis of the boundary value,
It is compared according to the timing inverted order of preceding N frame, obtains first optic disk centroid position and be less than or wait with average centroid position distance
In the boundary value frame as the initial frame.
The value range of the predetermined number Q in one of the embodiments, are as follows: N*10%≤Q≤N*30%.
It is described in one of the embodiments, to work as according to the variance being calculated and reference distance judgement
Whether the optic disk centroid position of preceding eye fundus image is noise, is included the following steps:
The variance and the reference distance are substituted into and determine formula: P (m)=a1×(dm-dBenchmark)+a2×lm, calculate
To corresponding decision content;Wherein, P (m) is the decision content being calculated, a1Coefficient, a are influenced for variance2On the basis of distance influence system
Number, dmFor the variance, dBenchmarkFor benchmark variance, lmFor the reference distance;
Judge whether the decision content is greater than or equal to preset value;
When judging that the decision content is greater than or equal to the preset value, then the optic disk of the current eye fundus image is determined
Centroid position is noise;When judging that the decision content is less than the preset value, then the view of the current eye fundus image is determined
Disk centroid position is not noise.
Continue to capture next frame eye fundus image in one of the embodiments, and during being determined, further include as
Lower step:
Count the noise quantity currently determined;
When counting the corresponding optic disk centroid position of current continuous b frame eye fundus image and being noise, then cancel to described
The judgement of next frame eye fundus image, and reset judgement queue.
Correspondingly, based on the same inventive concept, the present invention also provides a kind of optic disk tracing systems, including optic disk mass center position
It sets computing module, determine queue determining module, critical parameter computing module and noise judging treatmenting module;
Wherein, the optic disk centroid position computing module calculates institute for obtaining collected near-infrared fundus video stream
State the optic disk centroid position of each frame eye fundus image in near-infrared fundus video stream;
The judgement queue determining module, for extract in the near-infrared video stream initial preceding N frame eye fundus image and
Its optic disk centroid position, and according to the optic disk centroid position of extraction by determining current judgement in the preceding N frame eye fundus image
The initial frame of queue;Wherein, the current regular length threshold value for determining queue is n;
The critical parameter computing module, for for the current present frame eyeground figure for determining newly to capture in queue
Picture calculates the side of the optic disk centroid position of the present frame eye fundus image and the optic disk centroid position of previous valid frame eye fundus image
The distance between the optic disk centroid position of the optic disk centroid position of difference and the present frame eye fundus image and benchmark optic disk;Its
In, on the basis of the distance between optic disk centroid position of the optic disk centroid position of the present frame eye fundus image and benchmark optic disk away from
From;
The noise judging treatmenting module, for according to the variance being calculated and reference distance judgement
Whether the optic disk centroid position of present frame eye fundus image is noise;And in the optic disk mass center for determining the present frame eye fundus image
When position is noise, the present frame eye fundus image is rejected;Determine the optic disk centroid position of the present frame eye fundus image not
When being noise, the present frame eye fundus image and its optic disk centroid position are cached to output queue;
The noise judging treatmenting module is also used to when the rejecting or caching process for having carried out the present frame eye fundus image
Afterwards, the noise judgment module is jumped to, next frame in the current judgement queue is continued by the noise judgment module
The capture and judgement of eye fundus image, until being caught after the current length for determining queue reaches the fixed threshold n for subsequent
The each frame eye fundus image obtained, according to the principle of first in first out, is eliminated described current into while the current judgement queue
Determine the frame of queue head.
The judgement queue determining module includes extracting sub-module, mean value calculation submodule in one of the embodiments,
Block, average distance computational submodule, sorting sub-module and selection Comparative sub-module;
Wherein, extracting sub-module, for by extracting whole eye fundus images and its correspondence in the near-infrared fundus video stream
The optic disk centroid position;
The mean value calculation submodule, abscissa and ordinate for the optic disk centroid position described in top n respectively
Mean value calculation is carried out, average centroid position is obtained;
The average distance computational submodule, for calculating each described optic disk matter in N number of optic disk centroid position
The distance between heart position and the average centroid position, obtain N number of average distance;
The sorting sub-module, for carrying out size sequence to N number of average distance, and according to sequence from big to small
After the maximum average range for rejecting predetermined number, N-Q reasonable average distances are obtained;Wherein, Q is predetermined number;
The selection Comparative sub-module, for choosing the maximum average range in the reasonable average distance as boundary
Value is compared and on the basis of the boundary value according to the timing inverted order of preceding N frame, obtain first optic disk centroid position with
Average centroid position distance is less than or equal to the frame of the boundary value as the initial frame.
The noise judging treatmenting module includes that numerical value substitutes into submodule and decision content judgement in one of the embodiments,
Submodule;
Wherein, the numerical value substitutes into submodule, determines formula: P (m) for substituting into the variance and the reference distance
=a1×(dm-dBenchmark)+a2×lm, corresponding decision content is calculated;Wherein, P (m) is the decision content being calculated, a1For variance
Influence coefficient, a2On the basis of distance influence coefficient, dmFor the variance, dBenchmarkFor benchmark variance, lmFor the reference distance;
The decision content judging submodule, for judging whether the decision content is greater than or equal to preset value;
The decision content judging submodule is also used to when judging that the decision content is greater than or equal to the preset value,
The optic disk centroid position for then determining the current eye fundus image is noise;When judge the decision content be less than the preset value
When, then determine that the optic disk centroid position of the current eye fundus image is not noise.
It in one of the embodiments, further include noise quantity statistics module;
The noise quantity statistics module continues the current judgement queue for the noise judging treatmenting module
Whether the optic disk centroid position of middle next frame eye fundus image is to count the noise number currently determined in the decision process of noise
Amount;
The noise quantity statistics module is also used to count the corresponding optic disk mass center of current continuous b frame eye fundus image
When position is noise, then cancels the judgement to next frame eye fundus image, and jump to and determine queue determining module, sentenced by described
Determine the resetting of queue determining module and determines queue.
Correspondingly, based on the same inventive concept, the present invention also provides a kind of eyeground acquisition devices, including as above any institute
The optic disk tracing system stated.
Above-mentioned optic disk method for tracing calculates near-infrared fundus video stream by using near-infrared eyeground optic disk location algorithm
In each frame eye fundus image optic disk centroid position after, then by after determination determine queue in each frame eye fundus image and
Its optic disk centroid position is made whether the judgement for noise, wherein current using calculating when being made whether the judgement for noise
The variance of the optic disk centroid position of the optic disk centroid position of eye fundus image and its previous valid frame eye fundus image and current eyeground
Reference distance between the optic disk centroid position of image and the optic disk centroid position of benchmark optic disk, and according to the variance being calculated
Determined with reference distance, effectively eliminate the noise interference in output sequence, thus realize dynamic, near-infrared at
It is uncomfortable finally to efficiently solve traditional optic disk recognition positioning method and device for the tracing and positioning of optic disk position in the eyeground figure of picture
The problem of for dynamic, near infrared imaging eyeground figure.
Detailed description of the invention
Fig. 1 is the flow chart of a specific embodiment of optic disk method for tracing of the invention;
Fig. 2 is the process schematic of the another specific embodiment of optic disk method for tracing of the invention;
Fig. 3 is the structural schematic diagram of a specific embodiment of optic disk tracing system of the invention;
Fig. 4 is the structural schematic diagram of the another specific embodiment of optic disk tracing system of the invention.
Specific embodiment
To keep technical solution of the present invention clearer, the present invention is done further specifically below in conjunction with specific embodiment
It is bright.Wherein, it should be noted that the various details including helping to understand are described below, but these details will be considered as
It is merely exemplary.Therefore, it will be appreciated by those of ordinary skill in the art that in the situation for not departing from the scope of the present disclosure and spirit
Under, various embodiments described herein can be made various changes and modifications.In addition, for clarity and conciseness, known function and
The description of construction can be omitted.
Term used in following description and claims and vocabulary are not limited to document meaning, but only by inventor
For enabling the disclosure clearly and consistently to be understood.Therefore, to those skilled in the art it should be apparent that mentioning
For the description below to the various embodiments of the disclosure merely to exemplary purpose, rather than limit by appended claims and its
The purpose for the disclosure that equivalent limits.
It should be understood that unless the clear in addition instruction of context, otherwise singular also includes plural.Thus, for example,
Reference to " assembly surface " includes the reference to one or more such surfaces.
First include step S100, obtain as a specific embodiment of optic disk method for tracing of the invention referring to Fig. 1
Collected near-infrared fundus video stream is taken, and calculates the optic disk mass center position of each frame eye fundus image in near-infrared fundus video stream
It sets.Herein, it should be noted that since collected near-infrared fundus video stream includes several frame eye fundus images, adopting
After collecting near-infrared fundus video stream, need to carry out optic disk location tracking to each frame eye fundus image in video flowing.Fixed
In the tracing process of position, it is necessary first to determine the optic disk centroid position of each frame eye fundus image.Preferably, near-infrared optic disk can be used
Location algorithm calculates the optic disk centroid position of each frame eye fundus image.
Specifically, near-infrared optic disk location algorithm can be realized by following steps.Firstly, being extracted from near-infrared eyeground figure
The channel g of rgb.Then, using round lookup algorithm of increasing income, the eye fundus image in the channel g is handled, the multiple of extraction are obtained
Circular mass center, perimeter, the parameters such as radius may be obtained.In turn, it further according to the value of perimeter, radius, is compared with reasonable threshold value,
1 optimal solution is filtered out, optic disk mass center is obtained.
After being calculated by optic disk centroid position of the above method to each frame eye fundus image in video flowing
Step S200 is executed, the optic disk of preceding N frame eye fundus image and each frame eye fundus image initial in near-infrared fundus video stream is extracted
Centroid position, and currently sentenced according to the optic disk centroid position of each frame eye fundus image of extraction by determination in preceding N frame eye fundus image
Determine the initial frame of queue.
Specifically, by extracting initial preceding N frame eye fundus image and corresponding N number of optic disk centroid position in video flowing, and
When according to the optic disk centroid position of extraction by determining current judgement queue in preceding N frame eye fundus image, pass through step S210 first, by
Optic disk mass center position corresponding to initial preceding N frame eye fundus image and each frame eye fundus image before being extracted in near-infrared fundus video stream
It sets.At this point, the data extracted include N frame eye fundus image and N number of optic disk centroid position.Wherein, N number of optic disk centroid position can
It is characterized in the form of two-dimensional coordinate.That is, optic disk centroid position corresponding to the preceding N frame eye fundus image extracted is distinguished
Are as follows: (x1, y1)、(x2, y2)、(x3, y3)、……、(xN, yN)。
After extracting initial preceding N frame eye fundus image and its corresponding several optic disk centroid positions, i.e., executable step
Rapid S220, the abscissa to the N number of optic disk centroid position extracted and ordinate carry out mean value calculation respectively, obtain phase
The average centroid position answered.That is, according to calculation formula:
Average centroid position: (x is calculatedIt is average, yIt is average)。
Average centroid position (the x of N frame eye fundus image before being obtained by above-mentioned calculation formulaIt is average, yIt is average) and then pass through step
Rapid S230 calculates separately the distance between each optic disk centroid position and average centroid position in N number of optic disk centroid position, from
And obtain N number of average distance.That is, according to distance calculation formula:It is calculated N number of
Average distance: d1 is average、d2 is average、d3 is average、……、dM is average、……、dN is average.Wherein, the value of m arrives N, d for 1M is averageCharacterize N number of optic disk
M-th of optic disk centroid position (x in centroid positionm, ym) and average centroid position (xIt is average, yIt is average) between average distance.
And then step S240 is executed again, size sequence is carried out to the N number of average distance being calculated, and according to from big to small
Sequence successively reject the maximum average range of predetermined number after, obtain N-Q reasonable average distances.That is, to the N being calculated
A average distance: d1 is average、d2 is average、d3 is average、……、dM is average、……、dN is averageSize sequence is carried out, such as: according to sequence from big to small
It is arranged successively, to obtain the gradually smaller average distance ordered series of numbers of a columns value.Then, successively according still further to sequence from big to small
The average distance of predetermined number Q in average distance ordered series of numbers (such as: the value of predetermined number Q is the 30% of N) is weeded out, thus will
Remaining (N-Q) a average distance not being removed retains as reasonable average distance.Further, herein it should be noted that
The value range of predetermined number Q are as follows: N*10%≤Q≤N*30%.
Then, step S250 is executed, chooses the maximum average range in reasonable average distance as boundary value, and with boundary
It on the basis of value, is compared according to the timing inverted order of preceding N frame, obtains first optic disk centroid position and average centroid position distance
Less than or equal to the frame of boundary value as the current initial frame for determining queue.
Herein, it should be noted that the current regular length threshold value for determining queue is n.Wherein, n is less than N-Q.
After determining the current initial frame for determining queue through the above steps, the present frame eyeground that can newly be captured
The optic disk centroid position of image whether be noise judgement.Wherein, referring to fig. 2, in a tool of optic disk method for tracing of the invention
It whether is that mainly to pass through calculating current for the judgment basis of noise to the optic disk centroid position of eye fundus image in body embodiment
The variance of the optic disk centroid position of the optic disk centroid position of the present frame eye fundus image of judgement and its previous valid frame eye fundus image,
And the variance being calculated and benchmark variance are subjected to reasonability and sentence ratio, and then further according to the optic disk mass center of present frame eye fundus image
The distance between position and benchmark optic disk centroid position carry out reasonable evaluation to realize.
Therefore, it is based on above-mentioned judgment basis, it is specific in the another specific embodiment of optic disk method for tracing of the invention
Include: step S310, calculates in current judgement queue, the optic disk centroid position of current eye fundus image and its previous valid frame eyeground
The variance of the optic disk centroid position of image.Specifically, calculating the formula of variance are as follows:
Wherein, xm, ym are respectively the abscissa and ordinate of the optic disk centroid position of present frame eye fundus image, xq, yqRespectively
For the abscissa and ordinate of the optic disk centroid position of previous valid frame eye fundus image.By by the optic disk of present frame eye fundus image
The optic disk centroid position of centroid position and previous valid frame eye fundus image, which substitutes into, calculates formula of variance, can be obtained, and calculates simply,
It is convenient to realize.
Meanwhile the optic disk centroid position of present frame eye fundus image and the optic disk of benchmark optic disk are also calculated by step S320
The distance between centroid position is used as reference distance.Herein, it is noted that the value range of benchmark optic disk centroid position
Are as follows: it take eye fundus image center of circle midpoint as 0 coordinate of xy axis, the diametral distance of optic disk is dp, then for left eye, benchmark optic disk mass center
Coordinate is (- dp, -1/2dp), and for right eye, benchmark optic disk center-of-mass coordinate is (dp, -1/2dp).
In turn, then by step S400, according to variance being calculated and reference distance determines the optic disk of current eye fundus image
Whether centroid position is noise.
Specifically, carrying out the view of current eye fundus image according to the variance and reference distance that are calculated in step S400
When whether disk centroid position is the judgement of noise, first by step S410, variance and reference distance is substituted into and determine formula: P
(m)=a1×(dm-dBenchmark)+a2×lm, corresponding decision content is calculated.Wherein, P (m) is the decision content being calculated.a1For
Variance influences coefficient, value range are as follows: 0.85~0.95, preferably 0.93.a2On the basis of distance influence coefficient, value model
It encloses are as follows: 0.05~0.15, preferably 0.07.dmFor the optic disk centroid position of m frame eye fundus image (that is, present frame eye fundus image)
Variance between the optic disk centroid position of previous valid frame eye fundus image.dBenchmarkFor preset benchmark variance, value model
It encloses and may be configured as 0.5~1.5 disc diameter, preferably 0.8 disc diameter.lmFor the optic disk mass center position of current eye fundus image
It sets the distance between the optic disk centroid position of benchmark optic disk (that is, reference distance), value is between 2~4 disc diameters,
Preferably 2.7 disc diameters.
After corresponding decision content P (m) is calculated, which can be judged.Specifically, executing step
Rapid S420, judges whether the decision content P (m) is greater than or equal to preset value.When judging that it is pre- that the decision content P (m) is greater than or equal to
If then determining that the optic disk centroid position of current eye fundus image is noise when value.When judge the decision content P (m) be less than preset value
When, then determine that the optic disk centroid position of current eye fundus image is not noise.Wherein, the value of preset value is pre-set,
Preferably 0.9 disc diameter.
As a result, by by the optic disk matter of the optic disk centroid position of present frame eye fundus image and its previous valid frame eye fundus image
Between the variance of heart position and the optic disk centroid position of present frame eye fundus image and the optic disk centroid position of benchmark optic disk away from
From (as reference distance) this two parameters as noise judgment basis, can Effective selection go out the noise of optic disk positioning, and
The rejecting of noise is carried out in output sequence, thus ensured the opposite continuity and reasonability of optic disk label coordinate, it is final to realize
Effective tracking of optic disk.
Here, it should be pointed out that in the above-mentioned calculating variance the step of, the previous valid frame eye of present frame eye fundus image
Base map picture refers to before present frame eye fundus image, and a frame eye that is adjacent with present frame eye fundus image and not being noise
Base map picture.
Further, the another specific embodiment as optic disk method for tracing of the invention, when through the above steps to working as
Whether preceding eye fundus image is that noise determines after completing, and can carry out step S500, is continued under capturing in current judgement queue
One frame eye fundus image, and the judgement next frame eye fundus image captured being made whether for noise.
Wherein, next frame eye fundus image is being carried out whether in the decision process for being noise, it is preferred that it further includes step
S500 ' counts the noise quantity currently determined.When the optic disk centroid position for counting current continuous b frame eye fundus image is
When noise, then show currently to determine that queue is unreasonable determined by front, may be because the reasons such as shooting light leakage cause into
Image quality amount is bad and can not be judged, thus at this time then cancel to next frame eye fundus image whether be noise judgement, go forward side by side
Row determines the resetting of queue.Wherein, the value of b is preferably 5.
Further, after currently determining that the length of queue reaches threshold value n, for each frame eye fundus image newly captured,
It needs according to first in first out, determines that queue head eliminates a time earliest frame eye fundus image from current, until will acquire
To near-infrared fundus video stream in the optic disk centroid position of each frame eye fundus image carried out whether being that the judgement of noise is
Only.
Optic disk method for tracing of the invention as a result, by calculating each frame eye fundus image of near-infrared fundus video stream
Optic disk centroid position, and the optic disk centroid position of continuous N frame is cached by queue, to the continuous N frame optic disk centroid position of caching
The determination for determine the initial frame of queue, again to continuous several frame eyes of subsequent captured after determining the initial frame for determining queue
Base map picture carries out the calculating of variance and reference distance, to be weighed according to two parameters of the variance and reference distance being calculated
Re-computation carries out the screening and rejecting of noise according to weight, has ensured that the optic disk label coordinate of output is relatively continuous, reasonable, real
The purpose of optic disk tracking is showed.
Correspondingly, the present invention also provides a kind of optic disk tracing systems in order to realize any of the above-described kind of optic disk method for tracing.
Since the working principle of optic disk tracing system provided by the invention is identical as the principle of optic disk method for tracing provided by the invention or
It is similar, therefore overlaps will not be repeated.
A specific embodiment referring to Fig. 3, as optic disk tracing system 100 of the invention comprising optic disk centroid position
Computing module 110 determines queue determining module 120, critical parameter computing module 130 and noise judging treatmenting module 140.
Wherein, optic disk centroid position computing module 110 calculates close for obtaining collected near-infrared fundus video stream
The optic disk centroid position of each frame eye fundus image in infrared fundus video stream.Determine queue determining module 120, it is close red for extracting
Initial preceding N frame eye fundus image and its optic disk centroid position in outer video flowing, and according to the optic disk centroid position of extraction by preceding N frame
The current initial frame for determining queue is determined in eye fundus image;Wherein, the regular length threshold value for currently determining queue is n, and n is less than
Or it is equal to N.Critical parameter computing module 130, for for the current present frame eye fundus image for determining newly to capture in queue, meter
The variance of the optic disk centroid position of present frame eye fundus image and the optic disk centroid position of previous valid frame eye fundus image is calculated, and is worked as
The optic disk centroid position of previous frame eye fundus image and the distance between the optic disk centroid position of benchmark optic disk;Wherein, present frame eyeground
Distance on the basis of the distance between the optic disk centroid position of image and the optic disk centroid position of benchmark optic disk.Noise judgement processing mould
Block 140, for determined according to the variance that is calculated and reference distance present frame eye fundus image optic disk centroid position whether be
Noise;And when the optic disk centroid position for determining present frame eye fundus image is noise, present frame eye fundus image is rejected;It determines
When the optic disk centroid position of present frame eye fundus image is not noise, present frame eye fundus image is cached.Noise judging treatmenting module
140, it is also used to noise judgment module be jumped to, by noise after having carried out the rejecting or caching process of present frame eye fundus image
Judgment module continues the capture and judgement that currently determine next frame eye fundus image in queue, until the current judgement queue
In length reach the fixed threshold n after, for each frame eye fundus image of subsequent captured, into the current judgement queue
While, according to the principle of first in first out, eliminate the current frame for determining queue head.
Preferably, in optic disk tracing system 100 of the invention, determine that queue determining module 120 specifically includes extraction
Module, mean value calculation submodule, average distance computational submodule, sorting sub-module and selection Comparative sub-module (are not shown in figure
Out).
Wherein, extracting sub-module, for by N frame eye fundus image and its corresponding N before being extracted in near-infrared fundus video stream
A optic disk centroid position.
Mean value calculation submodule, for respectively to top n optic disk centroid position abscissa and ordinate be averaged
Value calculates, and obtains average centroid position.
Average distance computational submodule, for calculating each optic disk centroid position in N number of optic disk centroid position and being averaged
The distance between centroid position obtains N number of average distance.
Sorting sub-module for carrying out size sequence to N number of average distance, and is rejected according to sequence from big to small and is preset
After the maximum average range of number, N-Q reasonable average distances are obtained;Wherein, Q is predetermined number.
Comparative sub-module is chosen, for choosing the maximum average range in reasonable average distance as boundary value, and with side
On the basis of dividing value, be compared according to the timing inverted order of preceding N frame, obtain first optic disk centroid position and average centroid position away from
From the frame less than or equal to the boundary value as the initial frame.
Further, referring to fig. 4, the another specific embodiment as optic disk tracing system 100 of the invention, noise judgement
Processing module 140 includes that numerical value substitutes into submodule 141 and decision content judging submodule 142.Wherein, numerical value substitutes into submodule 141,
Formula: P (m)=a is determined for substituting into variance and reference distance1×(dm-dBenchmark)+a2×lm, corresponding judgement is calculated
Value;Wherein, P (m) is the decision content being calculated, a1Coefficient, a are influenced for variance2On the basis of distance influence coefficient, dmFor variance,
dBenchmarkFor benchmark variance, lmOn the basis of distance.
Decision content judging submodule 142, for judging whether decision content is greater than or equal to preset value.Decision content judges submodule
Block 142 is also used to then determine the optic disk mass center position of present frame eye fundus image when judging that decision content is greater than or equal to preset value
It is set to noise;When judging that decision content is less than preset value, then determine that the optic disk centroid position of present frame eye fundus image is not made an uproar
Point.
It further, further include having noise number in the another specific embodiment of optic disk tracing system 100 of the invention
Measure statistical module (not shown).Wherein, noise quantity statistics module continues for noise judging treatmenting module 140
Whether the current optic disk centroid position for determining next frame eye fundus image in queue is to count current in the decision process of noise and determine
Noise quantity out.
Herein, it should be noted that noise quantity statistics module is also used to count current continuous b frame eye fundus image
When corresponding optic disk centroid position is noise, then cancel the judgement to next frame eye fundus image, and jumps to and determine that queue is true
Cover half block 120 is reset by judgement queue determining module 120 and determines queue.
Optic disk tracing system 100 of the invention, by the way that optic disk centroid position computing module 110 is arranged, by optic disk mass center position
The optic disk centroid position that computing module 110 calculates each frame eye fundus image of near-infrared fundus video stream is set, and is sentenced by setting
Determine queue determining module 120, the continuous N frame optic disk centroid position cached by 120 Duis of judgement queue determining module carries out judgement team
The determination of column, again by continuous in 130 pairs of critical parameter computing module determining current judgement queues after determining judgement queue
Several frame eye fundus images carry out the calculating of variance and reference distance, thus again by noise judging treatmenting module 140 according to calculating
Obtained two parameters of variance and reference distance carry out weight calculation, and the screening and rejecting of noise are carried out according to weight, is ensured
The optic disk label coordinate of output is relatively continuous, reasonable, realizes the purpose of optic disk tracking.
In addition, it is based on any of the above-described kind of optic disk tracing system 100, the present invention also provides a kind of eyeground acquisition device,
Including as above any optic disk tracing system 100.
It is by being applied to eyeground acquisition device for any of the above-described kind of optic disk tracing system, so that eyeground of the invention acquires
The near-infrared eyeground preview video stream of number of frames per second can be effectively treated in device, and carries out optic disk to wherein each frame eye fundus image
Positioning, denoising, label realize the optic disk tracking effect under near infrared light, photographer can be assisted to shoot optic disk position and met
The eye fundus image of judgment criteria.
In addition, it should also be noted that, each technical characteristic of embodiment described above can be combined arbitrarily, to make
Description is succinct, and combination not all possible to each technical characteristic in above-described embodiment is all described, as long as however, these
Contradiction is not present in the combination of technical characteristic, all should be considered as described in this specification.
The embodiments described above only express several embodiments of the present invention, and the description thereof is more specific and detailed, but simultaneously
Limitations on the scope of the patent of the present invention therefore cannot be interpreted as.It should be pointed out that for those of ordinary skill in the art
For, without departing from the inventive concept of the premise, various modifications and improvements can be made, these belong to guarantor of the invention
Protect range.Therefore, the scope of protection of the patent of the invention shall be subject to the appended claims.
Claims (10)
1. a kind of optic disk method for tracing, which comprises the steps of:
Collected near-infrared fundus video stream is obtained, the view of each frame eye fundus image in the near-infrared fundus video stream is calculated
Disk centroid position;
Preceding N frame eye fundus image and its optic disk centroid position initial in the near-infrared video stream are extracted, and according to the institute of extraction
Optic disk centroid position is stated by determining the current initial frame for determining queue in the preceding N frame eye fundus image;Wherein, the current judgement
The regular length threshold value of queue is n;
For the current present frame eye fundus image for determining newly to capture in queue, the view of the present frame eye fundus image is calculated
The variance of the optic disk centroid position of disk centroid position and previous valid frame eye fundus image and the view of the present frame eye fundus image
The distance between the optic disk centroid position of disk centroid position and benchmark optic disk;Wherein, the optic disk matter of the present frame eye fundus image
Distance on the basis of the distance between optic disk centroid position of heart position and benchmark optic disk;
The optic disk centroid position of the present frame eye fundus image is determined according to the variance and the reference distance that are calculated
It whether is noise;And when the optic disk centroid position for determining the present frame eye fundus image is noise, the present frame is rejected
Eye fundus image;When the optic disk centroid position for determining the present frame eye fundus image is not noise, the present frame eyeground is cached
Image and its optic disk centroid position;
Continue to capture next frame eye fundus image in the current judgement queue and is determined, until the current judgement queue
After length reaches threshold value n, then for each frame eye fundus image of subsequent captured, into while the current judgement queue, root
According to the principle of first in first out, the frame of queue head is eliminated.
2. optic disk method for tracing according to claim 1, which is characterized in that described to extract in the near-infrared video stream just
The preceding N frame eye fundus image and the optic disk centroid position to begin, and according to the optic disk centroid position of extraction by the preceding N frame eye
The current initial frame for determining queue is determined in base map picture, is included the following steps:
By N frame eye fundus image and its corresponding N number of optic disk centroid position before being extracted in the near-infrared fundus video stream;
The abscissa to N number of optic disk centroid position and ordinate carry out mean value calculation respectively, obtain average centroid position;
Calculate between each described optic disk centroid position in N number of optic disk centroid position and the average centroid position away from
From obtaining N number of average distance;
Size sequence is carried out to N number of average distance, and rejects the average distance of predetermined number according to sequence from big to small
Afterwards, N-Q reasonable average distances are obtained;Wherein, Q is predetermined number;
The maximum average range in the reasonable average distance is chosen as boundary value, and on the basis of the boundary value, according to
The timing inverted order of preceding N frame is compared, and obtains first optic disk centroid position and average centroid position distance is less than or equal to institute
The frame of boundary value is stated as the initial frame.
3. optic disk method for tracing according to claim 2, which is characterized in that the value range of the predetermined number Q are as follows: N*
10%≤Q≤N*30%.
4. optic disk method for tracing according to claim 1, which is characterized in that the variance that the basis is calculated and
The reference distance determines whether the optic disk centroid position of the current eye fundus image is noise, is included the following steps:
The variance and the reference distance are substituted into and determine formula: P (m)=a1×(dm-dBenchmark)+a2×lm, it is calculated corresponding
Decision content;Wherein, P (m) is the decision content being calculated, a1Coefficient, a are influenced for variance2On the basis of distance influence coefficient, dm
For the variance, dBenchmarkFor benchmark variance, lmFor the reference distance;
Judge whether the decision content is greater than or equal to preset value;
When judging that the decision content is greater than or equal to the preset value, then the optic disk mass center of the current eye fundus image is determined
Position is noise;When judging that the decision content is less than the preset value, then the optic disk matter of the current eye fundus image is determined
Heart position is not noise.
5. optic disk method for tracing according to any one of claims 1 to 4, which is characterized in that continue to capture next frame eyeground
Image and during being determined, further includes following steps:
Count the noise quantity currently determined;
When counting the corresponding optic disk centroid position of current continuous b frame eye fundus image and being noise, then cancel to described next
The judgement of frame eye fundus image, and reset judgement queue.
6. a kind of optic disk tracing system, which is characterized in that including optic disk centroid position computing module, determine queue determining module,
Critical parameter computing module and noise judging treatmenting module;
Wherein, the optic disk centroid position computing module calculates described close for obtaining collected near-infrared fundus video stream
The optic disk centroid position of each frame eye fundus image in infrared fundus video stream;
The judgement queue determining module, for extracting preceding N frame eye fundus image and its view initial in the near-infrared video stream
Disk centroid position, and according to the optic disk centroid position of extraction by determining current judgement queue in the preceding N frame eye fundus image
Initial frame;Wherein, the current regular length threshold value for determining queue is n;
The critical parameter computing module, for being directed to the current present frame eye fundus image for determining newly to capture in queue,
The variance of the optic disk centroid position of the present frame eye fundus image and the optic disk centroid position of previous valid frame eye fundus image is calculated,
And the distance between the optic disk centroid position of the optic disk centroid position of the present frame eye fundus image and benchmark optic disk;Wherein,
Distance on the basis of the distance between the optic disk centroid position of the present frame eye fundus image and the optic disk centroid position of benchmark optic disk;
The noise judging treatmenting module, it is described current for being determined according to the variance and the reference distance that are calculated
Whether the optic disk centroid position of frame eye fundus image is noise;And in the optic disk centroid position for determining the present frame eye fundus image
When for noise, the present frame eye fundus image is rejected;The optic disk centroid position for determining the present frame eye fundus image is not made an uproar
When point, the present frame eye fundus image and its optic disk centroid position are cached to output queue;
The noise judging treatmenting module is also used to after having carried out the rejecting or caching process of the present frame eye fundus image,
The noise judgment module is jumped to, next frame eyeground in the current judgement queue is continued by the noise judgment module
The capture and judgement of image, until after the current length for determining queue reaches the fixed threshold n, for subsequent captured
Each frame eye fundus image, according to the principle of first in first out, eliminates the current judgement into while the current judgement queue
The frame of queue head.
7. optic disk tracing system according to claim 6, which is characterized in that the judgement queue determining module includes extracting
Submodule, mean value calculation submodule, average distance computational submodule, sorting sub-module and selection Comparative sub-module;
Wherein, extracting sub-module, for by extracting whole eye fundus images and its corresponding institute in the near-infrared fundus video stream
State optic disk centroid position;
The mean value calculation submodule is carried out for the abscissa of optic disk centroid position described in top n respectively and ordinate
Mean value calculation obtains average centroid position;
The average distance computational submodule, for calculating each optic disk mass center position in N number of optic disk centroid position
The distance between described average centroid position is set, N number of average distance is obtained;
The sorting sub-module for carrying out size sequence to N number of average distance, and is rejected according to sequence from big to small
After the maximum average range of predetermined number, N-Q reasonable average distances are obtained;Wherein, Q is predetermined number;
The selection Comparative sub-module, for choosing the maximum average range in the reasonable average distance as boundary value, and
It on the basis of the boundary value, is compared according to the timing inverted order of preceding N frame, obtains first optic disk centroid position and average matter
Heart positional distance is less than or equal to the frame of the boundary value as the initial frame.
8. optic disk tracing system according to claim 6, which is characterized in that the noise judging treatmenting module includes numerical value
Substitute into submodule and decision content judging submodule;
Wherein, the numerical value substitutes into submodule, determines formula: P (m)=a for substituting into the variance and the reference distance1
×(dm-dBenchmark)+a2×lm, corresponding decision content is calculated;Wherein, P (m) is the decision content being calculated, a1For variance shadow
Ring coefficient, a2On the basis of distance influence coefficient, dmFor the variance, dBenchmarkFor benchmark variance, lmFor the reference distance;
The decision content judging submodule, for judging whether the decision content is greater than or equal to preset value;
The decision content judging submodule is also used to then sentence when judging that the decision content is greater than or equal to the preset value
The optic disk centroid position of the fixed current eye fundus image is noise;When judging that the decision content is less than the preset value, then
The optic disk centroid position for determining the current eye fundus image is not noise.
9. according to the described in any item optic disk tracing systems of claim 6 to 8, which is characterized in that further include noise quantity statistics
Module;
The noise quantity statistics module, under continuing in the current judgement queue for the noise judging treatmenting module
Whether the optic disk centroid position of one frame eye fundus image is to count the noise quantity currently determined in the decision process of noise;
The noise quantity statistics module is also used to count the corresponding optic disk centroid position of current continuous b frame eye fundus image
When being noise, then cancel the judgement to next frame eye fundus image, and jump to and determine queue determining module, by the judgement team
The resetting of column determining module determines queue.
10. a kind of eyeground acquisition device, which is characterized in that including the described in any item optic disk tracing systems of claim 6 to 9.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811485787.XA CN109658395B (en) | 2018-12-06 | 2018-12-06 | Optic disc tracking method and system and eye fundus collection device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811485787.XA CN109658395B (en) | 2018-12-06 | 2018-12-06 | Optic disc tracking method and system and eye fundus collection device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109658395A true CN109658395A (en) | 2019-04-19 |
CN109658395B CN109658395B (en) | 2022-09-09 |
Family
ID=66112351
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811485787.XA Active CN109658395B (en) | 2018-12-06 | 2018-12-06 | Optic disc tracking method and system and eye fundus collection device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109658395B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112686927A (en) * | 2020-12-31 | 2021-04-20 | 上海易维视科技有限公司 | Human eye position regression calculation method |
CN113011450A (en) * | 2019-12-04 | 2021-06-22 | 深圳硅基智能科技有限公司 | Training method, training device, recognition method and recognition system for glaucoma recognition |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140294235A1 (en) * | 2013-03-29 | 2014-10-02 | National University Corporation Hokkaido University | Fundus image processing apparatus, fundus image processing method, and recording medium |
CN107180421A (en) * | 2016-03-09 | 2017-09-19 | 中兴通讯股份有限公司 | A kind of eye fundus image lesion detection method and device |
CN107424128A (en) * | 2017-06-20 | 2017-12-01 | 南京泰立瑞信息科技有限公司 | A kind of method and system that picture rich in detail frame is extracted from the fundus video recorded |
CN108024707A (en) * | 2015-07-24 | 2018-05-11 | 伟伦公司 | Automatic eye fundus image capture systems |
WO2018116321A2 (en) * | 2016-12-21 | 2018-06-28 | Braviithi Technologies Private Limited | Retinal fundus image processing method |
-
2018
- 2018-12-06 CN CN201811485787.XA patent/CN109658395B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140294235A1 (en) * | 2013-03-29 | 2014-10-02 | National University Corporation Hokkaido University | Fundus image processing apparatus, fundus image processing method, and recording medium |
CN108024707A (en) * | 2015-07-24 | 2018-05-11 | 伟伦公司 | Automatic eye fundus image capture systems |
CN107180421A (en) * | 2016-03-09 | 2017-09-19 | 中兴通讯股份有限公司 | A kind of eye fundus image lesion detection method and device |
WO2018116321A2 (en) * | 2016-12-21 | 2018-06-28 | Braviithi Technologies Private Limited | Retinal fundus image processing method |
CN107424128A (en) * | 2017-06-20 | 2017-12-01 | 南京泰立瑞信息科技有限公司 | A kind of method and system that picture rich in detail frame is extracted from the fundus video recorded |
Non-Patent Citations (2)
Title |
---|
吴慧等: "基于置信度计算的快速眼底图像视盘定位", 《计算机辅助设计与图形学学报》 * |
邹北骥等: "彩色眼底图像视盘自动定位与分割", 《光学精密工程》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113011450A (en) * | 2019-12-04 | 2021-06-22 | 深圳硅基智能科技有限公司 | Training method, training device, recognition method and recognition system for glaucoma recognition |
CN113011450B (en) * | 2019-12-04 | 2023-04-07 | 深圳硅基智能科技有限公司 | Training method, training device, recognition method and recognition system for glaucoma recognition |
CN112686927A (en) * | 2020-12-31 | 2021-04-20 | 上海易维视科技有限公司 | Human eye position regression calculation method |
CN112686927B (en) * | 2020-12-31 | 2023-05-12 | 上海易维视科技有限公司 | Human eye position regression calculation method |
Also Published As
Publication number | Publication date |
---|---|
CN109658395B (en) | 2022-09-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107341473B (en) | Palm characteristic recognition method, palm characteristic identificating equipment and storage medium | |
CN105894574B (en) | A kind of binocular three-dimensional reconstruction method | |
CN104463159B (en) | A kind of image processing method and device for positioning iris | |
CN104036231B (en) | Trunk identification device and method, terminal image detection method, device | |
CN104809424B (en) | Method for realizing sight tracking based on iris characteristics | |
CN104156951B (en) | A kind of white blood cell detection method for BAL fluid smear | |
CN104240264B (en) | The height detection method and device of a kind of moving object | |
WO2017118285A1 (en) | Method for rapidly extracting central point of circular image | |
CN110837750B (en) | Face quality evaluation method and device | |
EP2888718A1 (en) | Methods and systems for automatic location of optic structures in an image of an eye, and for automatic retina cup-to-disc ratio computation | |
TWI512640B (en) | Method for auto-depicting trends in object contours | |
JP7057873B2 (en) | Skipping rope counting method | |
CN109658395A (en) | Optic disk method for tracing and system and eyeground acquisition device | |
CN109359577A (en) | A kind of Complex Background number detection system based on machine learning | |
CN109547677A (en) | Eye fundus image image pickup method and system and equipment | |
CN105005813A (en) | Insect pest analyzing and counting method and system | |
CN106846293A (en) | Image processing method and device | |
CN105260559B (en) | A kind of paper pulp fiber morphological parameters calculation method based on contour area and contour thinning | |
CN109241963A (en) | Blutpunkte intelligent identification Method in capsule gastroscope image based on Adaboost machine learning | |
CN110037651A (en) | The method of quality control and device of eye fundus image | |
CN105930811B (en) | Palm texture feature detection method based on image processing | |
CN106973290A (en) | A kind of camera module stain method of testing and device | |
CN111353326A (en) | In-vivo detection method based on multispectral face difference image | |
CN106951823B (en) | A kind of infrared image undercarriage automatic identifying method | |
CN106067020A (en) | The system and method for quick obtaining effective image under real-time scene |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20231220 Address after: Room 707, Unit 1, Building 1, Courtyard 1, Longyu Middle Street, Huilongguan Town, Changping District, Beijing, 100085 Patentee after: Beijing Zhizhen Health Technology Co.,Ltd. Address before: 100085 room 1210, Building 29, yard 9, anningzhuang West Road, Haidian District, Beijing Patentee before: Dai Liming |