CN117115899A - Method and device for identifying white-eye venation characteristics, computer storage medium and electronic equipment - Google Patents
Method and device for identifying white-eye venation characteristics, computer storage medium and electronic equipment Download PDFInfo
- Publication number
- CN117115899A CN117115899A CN202311103068.8A CN202311103068A CN117115899A CN 117115899 A CN117115899 A CN 117115899A CN 202311103068 A CN202311103068 A CN 202311103068A CN 117115899 A CN117115899 A CN 117115899A
- Authority
- CN
- China
- Prior art keywords
- human eye
- image
- vein
- eye image
- characteristic point
- 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
- 238000000034 method Methods 0.000 title claims abstract description 74
- 238000003860 storage Methods 0.000 title claims abstract description 13
- 241001469893 Oxyzygonectes dovii Species 0.000 title claims description 21
- 241000282414 Homo sapiens Species 0.000 claims abstract description 207
- 210000003462 vein Anatomy 0.000 claims abstract description 108
- 210000004204 blood vessel Anatomy 0.000 claims abstract description 18
- 210000001835 viscera Anatomy 0.000 claims abstract description 18
- 239000003086 colorant Substances 0.000 claims abstract description 11
- 238000005192 partition Methods 0.000 claims abstract description 9
- 238000004422 calculation algorithm Methods 0.000 claims description 36
- 238000012549 training Methods 0.000 claims description 24
- 238000004364 calculation method Methods 0.000 claims description 18
- 238000012545 processing Methods 0.000 claims description 16
- 239000011159 matrix material Substances 0.000 claims description 9
- 238000003708 edge detection Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 6
- 238000005520 cutting process Methods 0.000 claims description 6
- 238000012706 support-vector machine Methods 0.000 claims description 6
- 238000000638 solvent extraction Methods 0.000 claims description 5
- 238000005286 illumination Methods 0.000 claims description 4
- GNFTZDOKVXKIBK-UHFFFAOYSA-N 3-(2-methoxyethoxy)benzohydrazide Chemical compound COCCOC1=CC=CC(C(=O)NN)=C1 GNFTZDOKVXKIBK-UHFFFAOYSA-N 0.000 claims description 3
- 206010025421 Macule Diseases 0.000 claims description 3
- 206010027146 Melanoderma Diseases 0.000 claims description 3
- 230000007812 deficiency Effects 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 210000001747 pupil Anatomy 0.000 claims description 3
- SWGJCIMEBVHMTA-UHFFFAOYSA-K trisodium;6-oxido-4-sulfo-5-[(4-sulfonatonaphthalen-1-yl)diazenyl]naphthalene-2-sulfonate Chemical compound [Na+].[Na+].[Na+].C1=CC=C2C(N=NC3=C4C(=CC(=CC4=CC=C3O)S([O-])(=O)=O)S([O-])(=O)=O)=CC=C(S([O-])(=O)=O)C2=C1 SWGJCIMEBVHMTA-UHFFFAOYSA-K 0.000 claims description 3
- 210000000264 venule Anatomy 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 abstract description 11
- 230000003044 adaptive effect Effects 0.000 abstract description 5
- 239000003814 drug Substances 0.000 description 11
- 238000004458 analytical method Methods 0.000 description 8
- 238000004891 communication Methods 0.000 description 6
- 238000003745 diagnosis Methods 0.000 description 5
- 241001504424 Zosteropidae Species 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 210000005036 nerve Anatomy 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- 230000009278 visceral effect Effects 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 230000001668 ameliorated effect Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 230000011514 reflex Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002792 vascular Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/18—Eye characteristics, e.g. of the iris
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/14—Vascular patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/18—Eye characteristics, e.g. of the iris
- G06V40/193—Preprocessing; Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/18—Eye characteristics, e.g. of the iris
- G06V40/197—Matching; Classification
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Human Computer Interaction (AREA)
- Ophthalmology & Optometry (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Vascular Medicine (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses a method and a device for identifying the characteristics of the veins of eyes based on a human eye viscera partition method, a computer storage medium and electronic equipment. The method comprises the steps of obtaining a human eye image to be identified; inputting the human eye image to be identified into a first network structure model, and calculating the vein form in the human eye image 、 Characteristic point coordinate information of color and trend; classifying and sorting the characteristic point coordinate information to generate a characteristic point data unit, wherein the classifying and sorting is based on the corresponding relation between the characteristic points and the venation; calculating human eye vein features corresponding to the feature point data units through a second network structure model, wherein the human eye vein features comprise blood vessel image information of colors, thicknesses and spots; based on the characteristic point data unit and the human eye venation characteristicA first image including at least human eye context feature information is generated. The method for identifying the human eye venation features has the advantages of high identification precision, small adaptive data volume, rapid iteration in actual use, better universality and convenience in rapid production.
Description
Technical Field
The present invention relates to the field of human eye image recognition, and in particular, to a method and apparatus for recognizing white eye venation characteristics, a computer storage medium, and an electronic device.
Background
In the eye auxiliary information processing in recent years, how to automatically and accurately extract the effective features of the eye images based on the images for comparison is a difficulty in research in the field, is also a main aspect for limiting the application of eye analysis related instruments and equipment, and how to extract the Chinese medicine venation features from the eye image is a general technical problem in the technical field of the identification of the current eye analysis instruments and equipment. The existing traditional Chinese medicine eye information analysis system identification method is to carry out deep learning based on all pixels of a full-eye image, and construct an AI-based analysis model, and although the implementation is relatively easy, the image has too much irrelevant element information, and the untreated machine learning can generate great interference on a final result, so that the accuracy is poor; meanwhile, the mode has the defects of long production period and poor stability due to the fact that a large number of samples are needed because of excessive input elements, the machine learning is performed again for a long time after each sample modification is carried out due to the large number of samples, and the quick response to the problems cannot be carried out, so that the practical application is difficult and heavy, and the commercialized equipment cannot be formed.
Therefore, how to provide a method and a device with good stability, high recognition accuracy and fast response and providing human eye vein feature recognition is a technical problem that needs to be broken through in the field.
Disclosure of Invention
Based on the defects of the prior art, the invention provides a method for identifying the characteristics of the venation of the eyes based on the human eye viscera partition method. The above-mentioned disadvantages can be ameliorated.
As one aspect of the present invention, there is provided a method for identifying a vein feature of a human eye based on a zonation method of viscera of the human eye, wherein the identification method comprises,
acquiring a human eye image to be identified;
inputting the human eye image to be identified into a first network structure model, and calculating the vein form in the human eye image 、 Characteristic point coordinate information of color and trend;
classifying and sorting the characteristic point coordinate information to generate a characteristic point data unit, wherein the classifying and sorting is based on the corresponding relation between the characteristic points and the venation;
calculating human eye vein features corresponding to the feature point data units through a second network structure model, wherein the human eye vein features comprise blood vessel image information of colors, thicknesses and spots;
and generating a first image at least comprising human eye vein feature information based on the feature point data unit and the human eye vein feature.
Preferably, in the method for identifying a feature of white eye venation provided in the first aspect of the present invention, the method further includes calculating a type of human eye venation according to the feature of human eye venation, where the type of human eye venation is information of venation morphology, venation color and special venation corresponding to viscera of the human body, and the type of human eye venation includes partition information.
Preferably, the method for identifying a vein feature of white eyes according to the first aspect of the present invention is characterized in that the vein type of the human eyes at least includes: coarse root, curved, extended, separated, branched, raised, blurred, vertical, black and through pupil,
and/or bright red, purple red, dark red, black in red, yellow in red, light yellow, gray and dark gray,
and/or six special venules including straight line, root deficiency, latticed, black spot, macula, and cyan spot.
Preferably, in the method for identifying a feature of a white eye according to the first aspect of the present invention, the second mesh model is trained by using a support vector machine algorithm, and the feature point data unit is calculated by using the second mesh model, which specifically includes: and inputting the characteristic point data unit into the second mesh structure model, and calculating the human eye venation characteristics corresponding to the characteristic point data unit in the human eye image.
Preferably, the first aspect of the present invention provides a method for identifying a vein feature of a white eye, wherein the first mesh model is trained by the following method:
step a, randomly generating one or more simple graphs, wherein the simple graphs comprise one or more of quadrangles, triangles, line segments and cubes, marking the vertexes of the simple graphs as characteristic points, taking the simple graphs and corresponding characteristic point coordinates as input, training an initial model by using a MagicPoint algorithm,
step b, inputting the human eye image to be identified to the initial model trained by the MagicPoint algorithm, performing a second training to obtain a further model,
step c, the human eye image to be identified and the feature point coordinates obtained by calculation based on the advanced model of the human eye image to be identified are used as input, and a MagicPoint algorithm is used for carrying out third-round training to obtain a high-order model;
d, taking the human eye image to be identified as input, and calculating the characteristic point coordinates corresponding to the human eye image to be identified by using the high-order model;
step e, manually checking the accuracy of the feature points extracted in the step d, if so, carrying out the step g, and if not, repeating the step c to iterate the high-order model;
and g, taking the eye image to be identified and the corresponding characteristic point coordinates as input, and training by using a SuperPoint algorithm to generate a final model.
Preferably, the method for identifying the vein features of the white eyes provided by the first aspect of the invention, wherein the second mesh model is trained by the following steps:
step A, carrying out normalization processing on each characteristic point data unit in a training set, converting the characteristic point data units into a 4x32 two-dimensional matrix, wherein each column in the matrix corresponds to each line segment of the characteristic point data unit, 4 data of each column are sequentially the vector angle, the vector length, the width and the color of one line segment of the characteristic point data unit, 32 data units can store data of 32 line segments at most, redundant data of more than 32 line segments are abandoned in the two-dimensional matrix, and less than 32 line segments are filled with 0;
step B, receiving the human eye venation characteristics marked by the manually-marked characteristic point data unit;
and C, taking the plurality of characteristic point data unit data normalized in the step A and the human eye vein type of each vein characteristic data unit marked in the step B as inputs, and training a model by using a support vector machine algorithm.
Preferably, the method for identifying a white eye vein feature provided in the first aspect of the present invention, the obtaining a human eye image to be identified specifically includes:
leading in a human eye image, wherein the human eye image is obtained by shooting by a camera,
removing the background of the eye image, retaining the eye image,
removing the highlight region formed by the reflection illumination light source on the human eye image, repairing the removed region by adopting an expansion algorithm,
the Canny edge detection algorithm, SOBEL edge detection algorithm and expansion algorithm are adopted to identify and cut the image, only the white eye area is reserved,
and taking the human eye image which only remains the white eye area after cutting as the human eye image to be identified.
In another embodiment of the present invention, the present invention further provides a device for identifying a vein feature of an eye based on a method for dividing viscera of the eye of a person, which specifically comprises:
an acquisition unit that acquires an eye image to be recognized;
the first computing unit inputs the human eye image to be identified into a first net structure model to compute the human eye imageVein morphology 、 Characteristic point coordinate information of color and trend;
the first generation unit is used for carrying out classification and sorting on the coordinate information of the feature points to generate a feature point data unit, wherein the classification and sorting are based on the corresponding relation between the feature points and the venation;
the second calculation unit is used for calculating the human eye vein features corresponding to the feature point data unit through a second net-shaped structure model, wherein the human eye vein features comprise blood vessel image information of colors, thicknesses and spots;
and the second generation unit is used for generating a first image at least comprising human eye vein feature information based on the feature point data unit and the human eye vein feature.
In yet another embodiment of the present invention, a computer storage medium is provided, the computer storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform the method steps of any of the above-described method steps for eye vein feature recognition based on human eye visceral partitioning. .
In another embodiment of the present invention, the present invention further provides an electronic device, including: a processor and a memory; wherein the memory stores a computer program adapted to be loaded by a processor and to perform the eye vein feature recognition method steps based on the eye visceral partitioning method as any one of the above.
The method for identifying the vein features of the eyes based on the human eye viscera partition method has the advantages of high identification precision, small adaptive data volume, rapid iteration in actual use, better universality and convenience in rapid production, and provides a reliable and stable method for identifying the vein features of the eyes for the application of eye analysis equipment of traditional Chinese medicine.
According to the invention, the complex and changeable human eye vein feature related information is orderly classified by decomposing the human eye vein feature recognition into feature point coordinate information calculation, classifying and sorting based on the corresponding relation between feature points and veins, and classifying and matching the human eye vein features, and the calculation recognition at different stages is respectively completed by two models, so that the data is ensured to be simple, the stability of data processing is further realized, the modularized thinking also realizes that samples are controllable, the self-adaptive updating is convenient, the quick recognition and the convenient production are realized, the problems of low recognition precision, high sample number requirement and slow sample training leading to production and long iteration period of the end-to-end deep learning of the current human eye image processing are solved, and the basic guarantee is provided for the accurate recognition and the wide application of the traditional Chinese medicine eye diagnosis.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims thereof as well as the appended drawings.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention and do not constitute a limitation on the invention. In the drawings:
FIG. 1 is a flowchart of a method for identifying the characteristics of the veins of eyes based on the zonation of the viscera of a human eye according to embodiment 1 of the present invention
FIG. 2 is a flowchart of a training method for a first mesh model according to embodiment 1 of the present invention
FIG. 3 is a flowchart of a training method of a second mesh model in embodiment 1 of the invention
FIG. 4 is a schematic diagram showing a device for identifying the venation characteristics of eyes in example 2 of the present invention
Fig. 5 is a schematic diagram of an electronic device in embodiment 3 of the invention
Detailed Description
The preferred embodiments of the present invention will be described below with reference to the accompanying drawings of the specification, it being understood that the preferred embodiments described herein are for illustration and explanation only, and not for limitation of the present invention, and embodiments of the present invention and features of the embodiments may be combined with each other without conflict.
Aiming at the problems of large data interference, poor accuracy, poor stability, incapability of timely self-adaptive updating according to actual use and untimely response and long production period in the field of human eye image recognition, the embodiment 1 of the invention provides a method for recognizing the vein features of eyes based on a human eye viscera partition method, which has the advantages of high recognition precision, small adaptive data volume, rapid iteration in actual use, better universality and convenience in quick production, and provides a reliable and stable method for recognizing the vein features of eyes for the application of eye analysis equipment of traditional Chinese medicine.
Specifically, in embodiment 1 of the present invention, a method for identifying the vein features of eyes based on the human eye viscera partition method specifically includes:
step 101: an image of the human eye to be identified is acquired,
specifically, the human eye image may be an auxiliary support for fixing eyes of different people and presetting a human eye state by an auxiliary device, a photographing device is used for photographing a target area of eyes of a human body to obtain the human eye image, and in order to support subsequent image processing and algorithm model identification, in the embodiment of the invention, a camera capable of clearly photographing the periphery of the eyes including details and colors in an eye body is preferably realized, and a photographing device with corresponding high pixels and resolution, such as a single-lens reflex camera, a smart phone camera or an industrial camera, can be selected.
The human eye image to be identified in the embodiment of the invention is a human eye image obtained by photographing a camera and performing image data processing on the human eye image. Specifically, the acquiring the eye image to be identified specifically includes:
leading in a human eye image, wherein the human eye image is obtained by shooting by a camera,
removing the background of the eye image, only retaining the eye image,
removing the highlight region formed by the reflective illumination source on the human eye image, repairing the removed region by adopting an improved expansion algorithm,
the image is identified and cut by adopting a Canny edge detection algorithm, a SOBEL edge detection algorithm and an improved expansion algorithm, only a white eye region is reserved,
and taking the human eye image which only remains the white eye area after cutting as the human eye image to be identified.
After the human eye image shot by the camera is processed through the steps, the subsequent eye area to be identified can be focused by cutting to the peripheral area, and meanwhile, image influence factors caused to the subsequent eye area to be identified due to factors such as reflected light, background, edge blurring and the like are eliminated, so that high-quality human eye images can be provided for subsequent accurate processing, and the accuracy of the subsequent overall human eye vein feature identification is improved.
Step 102: inputting the human eye image to be identified into a first network structure model, and calculating the vein form in the human eye image 、 And the color and trend feature point coordinate information.
Specifically, in the embodiment of the invention, the characteristic points are points capable of representing the characteristics of the image shape, specifically, 4 vertexes of a square, 3 vertexes of a triangle, 8 vertexes of a cube, inflection points of a broken line and crossing points of a crossing line are selected, and the points are connected to obtain the outline of the image. Because the shape and trend of the human eyes are related to the shape characteristics of the image, the key points can be used for accurate expression.
The feature point coordinate information calculated in embodiment 1 of the present invention may be scattered feature point coordinates that do not correspond to the context state.
Specifically, as shown in fig. 2, the first mesh structure model in the embodiment of the present invention is trained by the following method:
step a, randomly generating one or more simple graphs, wherein the simple graphs comprise one or more of quadrangles, triangles, line segments and cubes, marking the vertexes of the simple graphs as characteristic points, taking the simple graphs and corresponding characteristic point coordinates as input, training an initial model by using a MagicPoint algorithm,
step b, inputting the human eye image to be identified to the initial model trained by the MagicPoint algorithm, performing a second training to obtain a further model,
step c, the human eye image to be identified and the feature point coordinates obtained by calculation based on the advanced model of the human eye image to be identified are used as input, and a MagicPoint algorithm is used for carrying out third-round training to obtain a high-order model;
d, taking the human eye image to be identified as input, and calculating the characteristic point coordinates corresponding to the human eye image to be identified by using the high-order model;
step e, manually checking the accuracy of the feature points extracted in the step d, if so, carrying out the step g, and if not, repeating the step c to iterate the high-order model;
and g, taking the eye image to be identified and the corresponding characteristic point coordinates as input, and training by using a SuperPoint algorithm to generate a final model.
The invention sets the characteristic point coordinate information calculation model independently, rapidly and accurately calculates and extracts the characteristic points in the human eye image information to be identified through the first step, provides a basis for subsequent rapid and accurate processing in the form of the characteristic point coordinate information, and simultaneously, compared with the direct calculation of complete vein features or vein types based on the human eye image to be identified, the calculation process of the characteristic points is simple in data type and rule to be calculated, small in data quantity, and has the advantages of rapid and accurate calculation and the like.
Step 103: and classifying and sorting the characteristic point coordinate information to generate a characteristic point data unit, wherein the classifying and sorting is based on the corresponding relation between the characteristic points and the venation.
It should be noted that, in the embodiment of the present invention, the context is human tissue information capable of transmitting human biological signals, including blood vessels and nerve lines in the white eyes of the human body, and is generally represented by the form of image information.
After calculating the feature point coordinate information capable of representing the vein shape and the trend, the obtained feature point coordinate information is scattered feature point coordinates which are not associated with the vein shape and the trend corresponding to each feature point coordinate information. The feature point coordinate information calculated by the first mesh structure model is scattered feature points, which are not associated with the venation, but are substantially representative of information contained in all venations, so that step 103 can achieve high efficiency in subsequent calculation of venation features by sorting the scattered feature points in correspondence with each of the venations.
Therefore, step 103 in the embodiment of the present invention is configured to further perform sorting based on the correspondence between feature points and the venation, screen out feature points that can be associated with the venation, exclude feature points that are not associated with actual blood vessel positions and shape orientations in the image, and generally referred to as miscalculating feature points in the process, and after miscalculating feature points are removed, step 103 further performs sorting of feature points associated with the venation positions and shape orientations, where the sorting is divided into data sets according to an area where each venation in the image to be identified is located, and each data set corresponds to a position of each venation in the image to be identified.
In the embodiment of the present invention, in the step 103, the feature point coordinate information may be classified and ordered by adopting a topology structure based on a blood vessel pattern in the prior art, for example, by using an adaptive pattern detector to detect, identify and classify the feature point coordinate information, or by using machine learning trained by big data, and the specific classification and ordering method is based on the rule of human blood vessels and nerves distribution by those skilled in the art, and the feature point coordinate set obtained by calculation may be selected and optimized by themselves.
Step 104: and calculating the human eye vein features corresponding to the characteristic point data units through a second network structure model, wherein the human eye vein features comprise blood vessel image information of colors, thicknesses and spots. In step 104 of the embodiment of the present invention, the second mesh structure model is trained by using a support vector machine algorithm, and the human eye venation feature corresponding to the feature point data unit is calculated by using the second mesh structure model, which specifically includes: and inputting the characteristic point data unit into the second mesh structure model, and calculating the human eye venation characteristics corresponding to the characteristic point data unit in the human eye image.
Specifically, as shown in fig. 2, the second mesh model in the present invention is trained by the following steps:
step A, carrying out normalization processing on each characteristic point data unit in a training set, converting the characteristic point data units into a 4x32 two-dimensional matrix, wherein each column in the matrix corresponds to each line segment of the characteristic point data unit, 4 data of each column are sequentially the vector angle, the vector length, the width and the color of one line segment of the characteristic point data unit, 32 data units can store data of 32 line segments at most, redundant data of more than 32 line segments are abandoned in the two-dimensional matrix, and less than 32 line segments are filled with 0;
step B, receiving the human eye venation characteristics marked by the manually-marked characteristic point data unit;
and C, taking the plurality of characteristic point data unit data normalized in the step A and the human eye vein type of each vein characteristic data unit marked in the step B as inputs, and training a model by using a support vector machine algorithm.
In the embodiment of the invention, the human eye vein features at least comprise blood vessel image information of colors, thicknesses and spots. In addition to the above information, other vascular and neural image information that can reflect the veins of the human eye can be included in the present invention.
In the embodiment 1 of the present invention, through steps 102-104, a human eye image to be identified is processed by a model and an algorithm to obtain human eye vein features including at least blood vessel and nerve information, where the human eye vein features include shape, trend, color, thickness and other features of human eye blood vessels, so as to reflect blood vessel distribution in human eyes, and meanwhile, in order to correlate and display different states of human eyes, specifically, the human eye vein types at least include: coarse root, curved, extended, separated, branched, raised, blurred, vertical, black and through pupil,
and/or bright red, purple red, dark red, black in red, yellow in red, light yellow, gray and dark gray,
and/or six special venules including straight line, root deficiency, latticed, black spot, macula, and cyan spot.
Step 105: and generating a first image at least comprising human eye vein feature information based on the feature point data unit and the human eye vein feature.
Based on the theory of classifying human eye states by traditional Chinese medicine, the human eye vein features are obtained according to the corresponding relation between the human eye vein features and the human eye vein types, so that the automatic machine-calculated human eye image information is converted into human eye state information, the automatic provision of human eye state analysis is ensured, the automatic, high-efficiency and accurate machine breakthroughs are achieved, and the automatic machine-calculated human eye state analysis device plays a very important role in the application and popularization of an eye diagnosis instrument.
The method of dividing the human eye into three points in eight areas by the theory of dividing the left and right eyes into eight areas by the eye division, namely, the method of dividing the left and right eyes into three points in eight areas, namely, the method of dividing the left and right eyes into three points in 2007CB512707 is based on the theory of teaching an eye needle from the ancestor Peng Jingshan of the well-known traditional Chinese medicine, which is a part of the subject of the national basic research and development plan (973 plan) of the traditional Chinese medicine field, and the correspondence between the vein features and viscera is used for identifying the eye information. The method for identifying the white-eye venation features in the embodiment of the invention is based on the application of the human eye viscera partition method in computer image processing.
The human eye white vein feature recognition method provided by the embodiment 1 of the invention has the advantages of high recognition precision, adaptive data volume, rapid iteration in actual use, better universality and convenience in rapid production, and provides a reliable and stable human eye white vein feature recognition method for application in the field of traditional Chinese medicine eye diagnosis.
According to the invention, the complex and changeable human eye vein feature related information is orderly classified by decomposing the human eye vein feature recognition into feature point coordinate information calculation, classifying and sorting based on the corresponding relation between feature points and veins, and classifying and matching the human eye vein features, and the calculation recognition at different stages is respectively completed by two models, so that the single accuracy of data is ensured, the stability of data processing is further realized, the modularized thinking also realizes that samples are controllable, the self-adaptive updating is convenient, the quick recognition and the convenient production are realized, the problems of low recognition precision, high sample number requirement and slow sample training leading to production and long iteration period in the current human eye image processing are solved, and the basic guarantee is provided for the accurate recognition and the wide application of the traditional Chinese medicine eye diagnosis.
As shown in fig. 4, a second embodiment of the present invention provides a device for identifying a vein type of human eyes, which specifically includes an obtaining unit 401 for obtaining an image of a human eye to be identified;
the first calculating unit 403 inputs the eye image to be identified into a first mesh structure model, and calculates the vein form in the eye image 、 Characteristic point coordinate information of color and trend;
the first generating unit 405 performs classification and sorting on the feature point coordinate information to generate a feature point data unit, where the classification and sorting is based on a correspondence between feature points and the context;
a second calculating unit 407, configured to calculate, by using a second mesh model, a human eye vein feature corresponding to the feature point data unit, where the human eye vein feature includes blood vessel image information of color, thickness, and speckle;
the second generation unit 409 generates a first image including at least human eye vein feature information based on the feature point data unit and the human eye vein feature.
The human eye image to be identified in the embodiment of the invention is a human eye image obtained by photographing a camera and performing image data processing on the human eye image. Specifically, the acquiring the eye image to be identified specifically includes:
leading in a human eye image, wherein the human eye image is obtained by shooting by a camera,
removing the background of the eye image, only retaining the eye image,
removing the highlight region formed by the reflective illumination source on the human eye image, repairing the removed region by adopting an improved expansion algorithm,
the image is identified and cut by adopting a Canny edge detection algorithm, a SOBEL edge detection algorithm and an improved expansion algorithm, only a white eye region is reserved,
and taking the human eye image which only remains the white eye area after cutting as the human eye image to be identified.
After the human eye image shot by the camera is processed through the steps, the subsequent eye area to be identified can be focused by cutting to the peripheral area, and meanwhile, image influence factors caused to the subsequent eye area to be identified due to factors such as reflected light, background, edge blurring and the like are eliminated, so that high-quality human eye images can be provided for subsequent accurate processing, and the accuracy of identifying the vein features of the whole human eyes is improved.
The composition and training manner of the first mesh model and the second mesh model in the embodiments of the present invention are the portions of the disclosure already fully disclosed in embodiment 1, and will not be repeated here.
The recognition device of the human eye vein feature in the embodiment 2 of the invention shown in fig. 4 corresponds to the recognition method of the human eye vein feature in the embodiment 1, and the complex and changeable human eye vein feature related information is orderly classified by decomposing the human eye vein feature recognition into feature point coordinate information calculation, classification and sorting based on the corresponding relation between feature points and veins, and the manner of human eye vein feature calculation, classification and matching of the human eye vein feature, and the calculation recognition of different stages is respectively completed by dividing two models, so that the data is simple, the stability of data processing is further realized, the modular thought is also realized, the recognition accuracy of end-to-end deep learning in the current human eye image processing is not high, the sample number requirement is high, the sample training is slow, the problem of production and long iteration period is solved, and innovative support is provided for the accurate recognition and wide application of the traditional Chinese medicine eye diagnosis.
Fig. 5 illustrates a schematic structural diagram of an electronic device in embodiment 3 of the present invention, as shown in fig. 5, the electronic device may include: processor 510, communication interface (Communications Interface) 520, memory 530, and communication bus 540, wherein processor 510, communication interface 520, memory 530 complete communication with each other through communication bus 540. Processor 510 may invoke logic instructions in memory 530 to perform a method of the present invention for eye vein feature identification based on human eye viscera partitioning, the method comprising: acquiring a human eye image to be identified;
inputting the human eye image to be identified into a first network structure model, and calculating the vein form in the human eye image 、 Characteristic point coordinate information of color and trend;
classifying and sorting the characteristic point coordinate information to generate a characteristic point data unit, wherein the classifying and sorting is based on the corresponding relation between the characteristic points and the venation;
calculating human eye vein features corresponding to the feature point data units through a second network structure model, wherein the human eye vein features comprise blood vessel image information of colors, thicknesses and spots;
and generating a first image at least comprising human eye vein feature information based on the feature point data unit and the human eye vein feature.
Further, the logic instructions in the memory 530 described above may be implemented in the form of software functional units and may be stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
In another aspect, the present invention also provides a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, are capable of performing the method for identifying the venation feature of the human eye based on the visceral partitioning method of the human eye provided in the above embodiments, the method comprising: acquiring a human eye image to be identified;
inputting the human eye image to be identified into a first network structure model, and calculating the characteristic point coordinate information of the vein form, the color and the trend in the human eye image;
classifying and sorting the characteristic point coordinate information to generate a characteristic point data unit, wherein the classifying and sorting is based on the corresponding relation between the characteristic points and the venation;
calculating human eye vein features corresponding to the feature point data units through a second network structure model, wherein the human eye vein features comprise blood vessel image information of colors, thicknesses and spots;
and generating a first image at least comprising human eye vein feature information based on the feature point data unit and the human eye vein feature.
In yet another aspect, the present invention provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, is implemented to perform the above provided methods for identifying white-eye context features based on human-eye viscera partition methods, the method comprising: acquiring a human eye image to be identified;
inputting the human eye image to be identified into a first network structure model, and calculating the vein form in the human eye image 、 Characteristic point coordinate information of color and trend;
classifying and sorting the characteristic point coordinate information to generate a characteristic point data unit, wherein the classifying and sorting is based on the corresponding relation between the characteristic points and the venation;
calculating human eye vein features corresponding to the feature point data units through a second network structure model, wherein the human eye vein features comprise blood vessel image information of colors, thicknesses and spots;
and generating a first image at least comprising human eye vein feature information based on the feature point data unit and the human eye vein feature.
The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. A method for identifying the characteristics of the veins of eyes based on the zonation method of viscera of eyes of human beings is characterized in that the identification method comprises the following steps,
acquiring a human eye image to be identified;
inputting the human eye image to be identified into a first network structure model, and calculating the vein form in the human eye image 、 Characteristic point coordinate information of color and trend;
classifying and sorting the characteristic point coordinate information to generate a characteristic point data unit, wherein the classifying and sorting is based on the corresponding relation between the characteristic points and the venation;
calculating human eye vein features corresponding to the feature point data units through a second network structure model, wherein the human eye vein features comprise blood vessel image information of colors, thicknesses and spots;
and generating a first image at least comprising human eye vein feature information based on the feature point data unit and the human eye vein feature.
2. The method of claim 1, further comprising calculating a human eye context type based on the human eye context feature, the human eye context type being information of context morphology, context color, and special context in correspondence with human viscera, the human eye context type including zonal information.
3. The method of identifying white-eye context features of claim 2, wherein the human eye context type comprises at least: ten forms of thick root, curved, stretched, separated, forked, raised, blurred, vertical, black circle and through pupil; and/or bright red, purple red, dark red, black in red, yellow in red, light yellow, gray and dark gray; and/or six special venules including straight line, root deficiency, latticed, black spot, macula, and cyan spot.
4. The method for identifying the features of the eyes as claimed in claim 1, wherein the second mesh model is trained by a support vector machine algorithm, and the features of the eyes corresponding to the feature point data unit are calculated by the second mesh model, specifically comprising: and inputting the characteristic point data unit into the second mesh structure model, and calculating the human eye venation characteristics corresponding to the characteristic point data unit in the human eye image.
5. The method of identifying white-eye context features of claim 1, wherein the first mesh model is trained by:
step a, randomly generating one or more simple graphs, wherein the simple graphs comprise one or more of quadrangles, triangles, line segments and cubes, marking the vertexes of the simple graphs as characteristic points, taking the simple graphs and corresponding characteristic point coordinates as input, training an initial model by using a MagicPoint algorithm,
step b, inputting the human eye image to be identified to the initial model trained by the MagicPoint algorithm, performing a second training to obtain a further model,
step c, the human eye image to be identified and the feature point coordinates obtained by calculation based on the advanced model of the human eye image to be identified are used as input, and a MagicPoint algorithm is used for carrying out third-round training to obtain a high-order model;
d, taking the human eye image to be identified as input, and calculating the characteristic point coordinates corresponding to the human eye image to be identified by using the high-order model;
step e, manually checking the accuracy of the feature points extracted in the step d, if so, carrying out the step g, and if not, repeating the step c to iterate the high-order model;
and g, taking the eye image to be identified and the corresponding characteristic point coordinates as input, and training by using a SuperPoint algorithm to generate a final model.
6. The method of white-eye vein feature recognition as set forth in claim 1 wherein said second mesh model is trained by the steps of:
step A, carrying out normalization processing on each characteristic point data unit in a training set, converting the characteristic point data units into a 4x32 two-dimensional matrix, wherein each column in the matrix corresponds to each line segment of the characteristic point data unit, 4 data of each column are sequentially the vector angle, the vector length, the width and the color of one line segment of the characteristic point data unit, redundant data of more than 32 line segments are discarded in the two-dimensional matrix, and less than 32 line segments are filled with 0;
step B, receiving the human eye venation characteristics marked by the manually-marked characteristic point data unit;
and C, taking the plurality of characteristic point data unit data normalized in the step A and the human eye vein type of each vein characteristic data unit marked in the step B as inputs, and training a model by using a support vector machine algorithm.
7. The method for identifying white-eye context features as defined in claim 1, wherein said obtaining an image of a human eye to be identified comprises:
leading in a human eye image, wherein the human eye image is obtained by shooting by a camera,
removing the background of the eye image, retaining the eye image,
removing the highlight region formed by the reflection illumination light source on the human eye image, repairing the removed region by adopting an expansion algorithm,
the Canny edge detection algorithm, SOBEL edge detection algorithm and expansion algorithm are adopted to identify and cut the image, only the white eye area is reserved,
and taking the human eye image which only remains the white eye area after cutting as the human eye image to be identified.
8. An eye vein feature recognition device based on human eye viscera partition method, which is characterized by comprising the following steps:
an acquisition unit that acquires an eye image to be recognized;
the first calculating unit inputs the human eye image to be identified into a first net structure model to calculate the vein form in the human eye image 、 Characteristic point coordinate information of color and trend;
the first generation unit is used for carrying out classification and sorting on the coordinate information of the feature points to generate a feature point data unit, wherein the classification and sorting are based on the corresponding relation between the feature points and the venation;
the second calculation unit is used for calculating the human eye vein features corresponding to the feature point data unit through a second net-shaped structure model, wherein the human eye vein features comprise blood vessel image information of colors, thicknesses and spots;
and the second generation unit is used for generating a first image at least comprising human eye vein feature information based on the feature point data unit and the human eye vein feature.
9. A computer storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform the eye vein feature recognition method steps of any one of claims 1-7 based on the human eye viscera partitioning method.
10. An electronic device, comprising: a processor and a memory; wherein the memory stores a computer program adapted to be loaded by a processor and to perform the method steps of the method for identifying features of the eye according to any one of claims 1-7 based on the zonation of the viscera of the eye.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311103068.8A CN117115899A (en) | 2023-08-28 | 2023-08-28 | Method and device for identifying white-eye venation characteristics, computer storage medium and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311103068.8A CN117115899A (en) | 2023-08-28 | 2023-08-28 | Method and device for identifying white-eye venation characteristics, computer storage medium and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117115899A true CN117115899A (en) | 2023-11-24 |
Family
ID=88801785
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311103068.8A Pending CN117115899A (en) | 2023-08-28 | 2023-08-28 | Method and device for identifying white-eye venation characteristics, computer storage medium and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117115899A (en) |
-
2023
- 2023-08-28 CN CN202311103068.8A patent/CN117115899A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109815850B (en) | Iris image segmentation and positioning method, system and device based on deep learning | |
| CN105917353B (en) | Feature extraction and matching for biological identification and template renewal | |
| Neal et al. | Measuring shape | |
| CN111178197A (en) | Mass R-CNN and Soft-NMS fusion based group-fed adherent pig example segmentation method | |
| CN109782902A (en) | A kind of operation indicating method and glasses | |
| CN107657612A (en) | Suitable for full-automatic the retinal vessel analysis method and system of intelligent and portable equipment | |
| CN109117857B (en) | Biological attribute identification method, device and equipment | |
| CN109583364A (en) | Image-recognizing method and equipment | |
| CN110543848B (en) | Driver action recognition method and device based on three-dimensional convolutional neural network | |
| CN109034133A (en) | A kind of face identification method and device | |
| CN102567734A (en) | Specific value based retina thin blood vessel segmentation method | |
| CN109063720A (en) | Handwritten word training sample acquisition methods, device, computer equipment and storage medium | |
| CN109087310A (en) | Dividing method, system, storage medium and the intelligent terminal of Meibomian gland texture region | |
| CN110909601B (en) | Beautiful pupil identification method and system based on deep learning | |
| CN116051631A (en) | Light spot labeling method and system | |
| CN109919098A (en) | target object identification method and device | |
| CN113191189A (en) | Face living body detection method, terminal device and computer readable storage medium | |
| CN110428431B (en) | Method, device and equipment for segmenting cardiac medical image and storage medium | |
| CN115359548B (en) | Handheld intelligent pupil detection device and detection method | |
| CN117115899A (en) | Method and device for identifying white-eye venation characteristics, computer storage medium and electronic equipment | |
| CN113012030A (en) | Image splicing method, device and equipment | |
| CN110008881A (en) | The recognition methods of the milk cow behavior of multiple mobile object and device | |
| CN112801238B (en) | Image classification method and device, electronic equipment and storage medium | |
| CN114820537A (en) | Dry eye FBUT detection method and system based on deep learning and storage medium | |
| CN112949585A (en) | Identification method and device for blood vessels of fundus image, electronic equipment and storage medium |
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 |