CN115375611A - Model training-based refraction detection method and detection system - Google Patents
Model training-based refraction detection method and detection system Download PDFInfo
- Publication number
- CN115375611A CN115375611A CN202110563271.8A CN202110563271A CN115375611A CN 115375611 A CN115375611 A CN 115375611A CN 202110563271 A CN202110563271 A CN 202110563271A CN 115375611 A CN115375611 A CN 115375611A
- Authority
- CN
- China
- Prior art keywords
- sample set
- detection
- refraction
- pupil
- information
- 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
Images
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
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
- A61B3/0016—Operational features thereof
- A61B3/0025—Operational features thereof characterised by electronic signal processing, e.g. eye models
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
- A61B3/10—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions
- A61B3/103—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions for determining refraction, e.g. refractometers, skiascopes
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
- A61B3/10—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions
- A61B3/11—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions for measuring interpupillary distance or diameter of pupils
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
- A61B3/10—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions
- A61B3/14—Arrangements specially adapted for eye photography
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
Abstract
The invention discloses a refraction detection method and a detection system based on model training, wherein the method comprises the following steps: acquiring a pupil information sample set, a personnel information sample set and a refraction data set of a person to be acquired, wherein the pupil information sample set at least comprises a pupil image and pupil position information, and the personnel information sample set at least comprises environmental illumination information and personnel station position information; carrying out data preprocessing on the pupil information sample set and the personnel information sample set, and constructing a detection training sample set according to the refraction data set; and training the refraction detection model according to the detection training sample set so as to perform refraction detection on the person to be detected through the trained refraction detection model. The method solves the problems of low efficiency and low accuracy of the traditional refraction detection method.
Description
Technical Field
The invention relates to the technical field of refraction detection, in particular to a refraction detection method and a detection system based on model training.
Background
The eye is the sensory organ of the human body observing objective things. The light rays emitted or reflected from distant or near objects outside, whether parallel or dispersed, are refracted by the dioptric system of the eye, and then gathered and combined on the retina, and then the light rays are driven to transmit to the visual center of the brain through the visual path to generate vision. Ametropia refers to the condition that when the eye does not use adjustment, parallel rays cannot form a clear object image on the retina after passing through the refractive action of the eye, and the object image is formed in front of or behind the retina. Myopia and astigmatism are often said to be manifestations of refractive error. The incidence of myopia of children and teenagers in China increases year by year with the increase of age, the incidence still shows a rising trend, and the myopia prevention and control task is difficult. In order to be able to detect children and adolescents with vision problems as early as possible for timely and effective treatment and intervention, refractive testing is required. In the previous refraction detection method, a screening instrument is a common method, the screening instrument generally adopts an eccentric photography optometry method, and is suitable for large-range refraction detection, but the problems that the screening instrument is easy to be subjected to environmental factor images, the position matching requirement of a detected person is high, and the professional knowledge of a user is required to a certain extent exist, so that the detection accuracy of the traditional eccentric photography optometry method needs to be improved.
Disclosure of Invention
In view of this, the embodiment of the present invention provides a refractive detection method and a detection system based on model training, so as to improve the accuracy of refractive detection.
According to a first aspect, a model training based refraction detection method, the method comprising:
acquiring a pupil information sample set, a personnel information sample set and a refraction data set of a person to be acquired, wherein the pupil information sample set at least comprises a pupil image and pupil position information, and the personnel information sample set at least comprises environmental illumination information and personnel station position information;
carrying out data preprocessing on the pupil information sample set and the personnel information sample set, and constructing a detection training sample set according to the refraction data set;
and training the refraction detection model according to the detection training sample set so as to perform refraction detection on the person to be detected through the trained refraction detection model.
Optionally, the obtained pupil information sample set at least includes a pupil image and pupil position information, and includes:
positioning the eye region of the face image of the collected person to obtain the pupil position information;
and the pupil image is obtained by dividing the eye region.
Optionally, the preprocessing the data of the pupil information sample set and the personnel information sample set to construct a detection training sample set includes:
carrying out image preprocessing on the images in the pupil information sample set;
carrying out data quality pretreatment on the data in the personnel information sample set;
and marking the pupil information sample set and the personnel information sample set according to the refraction data set, and constructing the marked samples as the detection training sample set.
Optionally, the labeling the pupil information sample set and the staff information sample set according to the refraction data set includes:
the refractive data set includes refractive values including sphere, cylinder, and axis information.
Optionally, the personal information sample set further includes at least personal information of the person to be collected.
Optionally, a model training based refraction detection method, the method further comprising:
acquiring the personnel information sample set and the historical refraction data set;
performing data quality pretreatment on the personnel information sample set, and constructing a prediction training sample set according to a historical refraction data set;
and training the vision prediction model according to the prediction training sample set so as to predict the vision of the person to be detected through the trained vision prediction model.
According to a second aspect, a model-training based refractive detection system, the system comprising:
the data acquisition module is used for acquiring a pupil information sample set, a personnel information sample set and a refraction data set of a person to be acquired, wherein the pupil information sample set at least comprises a pupil image and pupil position information, and the personnel information sample set at least comprises environmental illumination information and personnel station position information;
the training sample module is used for carrying out data preprocessing on the pupil information sample set and the personnel information sample set and constructing a detection training sample set according to the refraction data set;
and the detection module trains the refraction detection model according to the model training sample set so as to perform refraction detection on the person to be detected through the trained refraction detection model.
Optionally, a model-based trained refractive detection system, the system comprising:
the data acquisition module is used for acquiring the personnel information sample set and the historical refraction data set;
the training sample module is used for carrying out data quality preprocessing on the personnel information sample set and constructing a prediction training sample set according to the historical refraction data set;
and the prediction module is used for training the vision prediction model according to the prediction training sample set so as to predict the vision of the person to be detected through the trained vision prediction model.
According to a third aspect, a refractive detection apparatus comprises:
a memory and a processor, the memory and the processor being communicatively coupled to each other, the memory having stored therein computer instructions, and the processor performing the method of the first aspect, or any one of the optional embodiments of the first aspect, by executing the computer instructions.
According to a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, where the computer-readable storage medium stores computer instructions for causing a computer to thereby execute the method described in the first aspect, or any one of the optional implementation manners of the first aspect.
The technical scheme of the invention has the following advantages:
the embodiment of the invention provides a model training-based refraction detection method, which specifically comprises the following steps: firstly, acquiring a pupil information sample set, a personnel information sample set and a refraction data set, wherein the pupil information sample set at least comprises a pupil image and pupil position information, and the personnel information sample set at least comprises environmental illumination information and personnel station position information; then, carrying out data preprocessing on the obtained pupil information sample set and the personnel information sample set, and constructing a detection training sample set according to the refraction data set; and finally, training the refraction detection model according to the detection training sample set so as to perform refraction detection on the person to be detected through the trained refraction detection model. The pupil images under a large number of different environmental factors are combined with personal information of the collected personnel, so that sample data with the situation covering most conditions is obtained, the refraction detection model constructed through algorithms such as machine learning is complete as much as possible, the refraction detection can be effectively carried out on most personnel to be detected, and the accuracy of the refraction detection is improved. And then, a vision prediction model is established according to the trend of the vision of the collected person along with the change of time through historical vision data modeling, so that the method can realize future vision prediction besides refraction detection. The practicability of the method is further improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a schematic diagram of steps of a model training-based refraction detection method according to an embodiment of the invention;
FIG. 2 is a schematic image processing flow chart of a refraction detection method based on model training according to an embodiment of the invention;
FIG. 3 is a schematic diagram of a model-based training refractive detection system according to an embodiment of the invention;
FIG. 4 is a schematic structural diagram of a model training-based refraction detection system according to an embodiment of the invention
FIG. 5 is a schematic view of a refractive detection apparatus according to an embodiment of the present invention;
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The technical features mentioned in the different embodiments of the invention described below can be combined with each other as long as they do not conflict with each other.
Referring to fig. 1, a refraction detection method based on model training provided by an embodiment of the present invention specifically includes the following steps:
step S101: the method comprises the steps of obtaining a pupil information sample set, a personnel information sample set and a refraction data set of a person to be collected, wherein the pupil information sample set at least comprises a pupil image and pupil position information, and the personnel information sample set at least comprises environment illumination information and personnel station position information.
Specifically, the traditional screening instrument generally adopts an eccentric photography optometry method, is suitable for large-range refraction detection, is easy to be subjected to environmental factor images, has higher requirements on position matching of a detected person, and has certain requirements on professional knowledge of a user. The principle of eccentric photography optometry is that near-infrared light-emitting diodes are adopted to form a light source array, light rays emit to a pupil to be detected at a certain distance at a specific angle, pass through the pupil and strike on the retina to be reflected by the retina, and during the period, the light rays are refracted by an eyeball dioptric system and then emit from a pupil area to be shot by an image acquisition device. Therefore, the refractive state and the adjustment state of the eye to be detected determine the shape and the brightness of the light shadow of the pupil area of the eye to be detected, and then the professional obtains the refractive information of the person to be detected through the position and the state of the crescent-shaped bright area in the pupil area. It is a good solution to construct a test model trained with a large amount of data to automatically perform the refraction test. Based on the method, firstly, a hardware system based on the eccentric photography optometry principle is used for acquiring a large number of facial images of an acquired person, wherein the facial images can be a front face and a side face with a certain rotation angle. And then extracting the pupil image and the pupil position in the face image to obtain a complete pupil information sample set, wherein the pupil information sample set may include the face image besides the pupil image and the pupil position information, and further may include the upper body image of the captured person, and the like, which is not limited by the present invention. Wherein the pupil location information is a location of the eye region in the face image. The personal information samples corresponding to the pupil images include, but are not limited to: the method comprises the steps of obtaining ambient illumination information of pupil images under different illumination conditions, and obtaining personnel station position information of stations in different directions at different distances from a photographing device. The refractive data set may be obtained by subjective refraction of the person at the point of vision detection, which may include sphere, cylinder, axis, interpupillary distance, and the like. The personnel station position information can be obtained by distance measurement through one or more sensors such as an ultrasonic distance measuring sensor, a laser distance measuring sensor, an infrared distance measuring sensor, a radar distance measuring sensor and the like by utilizing a monocular vision or binocular vision and an image processing algorithm through an image sensor. Ambient lighting information may be obtained by a lighting sensor. In particular, the set of person information samples may further include: the name, class, sex, age, height, weight, BMI index, parent myopia, historical vision data, historical refraction data, etc. of the person to be collected. The condition that most of refraction information appears can be established through a large amount of pupil image information and personnel information for it is more random to treat the personnel that detect posture, position in the testing process, thereby perfect data sample has constructed perfect detection model, treat that the personnel that detect carry out refraction detection and reduced the step that professional confirm refraction information by experience, make refraction detection more level and smooth to the threshold of personnel's quality, avoided personnel's subjective judgement, improved the refraction detection accuracy greatly.
Specifically, in one embodiment, the personnel information sample set further comprises at least personal information of the person to be collected. Personal information includes, but is not limited to: the name, class, sex, age, height, weight, BMI index, parent myopia, historical vision data, historical refractive data and the like of the person to be collected. Through more detailed personnel information, various objective factors are comprehensively considered so as to improve the accuracy and reliability of the refraction detection method.
Step S102: and carrying out data preprocessing on the pupil information sample set and the personnel information sample set, and constructing a detection training sample set according to the refraction data set. Specifically, after the pupil information sample set and the person information sample set are acquired, the pre-processing on the acquired image data mainly includes, but is not limited to, one or more of operations of image filtering, cropping, transforming, splicing and combining, and the like, where the splicing and combining may include operations of splicing image regions or synthesizing image multiple channels, and the like. The preprocessing of the acquired numerical data mainly comprises but is not limited to one or more of data cleaning, data integration, data specification, data transformation and the like. Through the processing, the characteristics in the images of the pupil information sample set are clearer, unqualified images are eliminated, the data quality is ensured, and the reliability of establishment of a subsequent training model is improved. And the unqualified related personnel information data are removed by the data preprocessing of the personnel information sample set, the data quality is also ensured, and the reliability of the establishment of a subsequent training model is improved. And then, labeling data by combining the personnel information sample set and the pupil information sample set to form a detection training sample set, wherein the label form of the detection training sample set is a refraction value in the refraction data set and comprises sphere lens, cylindrical lens and axis position information. The label source can be obtained by detection of professional medical institutions or detection of professional optometry equipment, and the detection training sample set is formed so as to establish a detection model in the following steps.
Step S103: and training the refraction detection model according to the detection training sample set so as to perform refraction detection on the person to be detected through the trained refraction detection model. Specifically, the test training sample set obtained in step S102 is used as an input of the refraction testing model, so that the relevant parameters in the refraction testing model are adjusted through iterative training to establish a complete refraction testing model. The data set is trained by at least one learning algorithm, and the parameters are adjusted to optimize the model. Learning algorithms herein include, but are not limited to: one or more of a neural network model, a regression model, a least squares model, a support vector machine, a Markov algorithm, etc., wherein the neural network model comprises one or more of a deep neural network, a recurrent neural network, a convolutional neural network, etc. In the embodiment of the invention, the CNN method is adopted to construct the detection model, and then the refraction detection model is established to detect the refraction of the person to be detected, so that the aim of automatic detection is fulfilled.
Specifically, in an embodiment, step S101 of the foregoing step specifically includes the following steps:
step S201: and positioning the eye region of the face image of the acquired person to obtain pupil position information. Specifically, as shown in fig. 2, after an image of the face of the subject is acquired by using a hardware system based on the principle of eccentric photography optometry, since pupil information of the person to be acquired is required for performing refraction detection, irrelevant image information needs to be rejected, and excessive sample noise is avoided. First the eye region is located in the facial image. In the embodiment of the present invention, a template matching method is used to position the eye region, a cascade classifier method may also be used to position the eye region, and the positioning of the eye region is performed according to the gray-level value relationship of the pixels of the image by xor, which is not limited in the present invention. From the positioned eye region image, pupil position information is acquired, and the eye region image is subjected to the processing of the next step S202.
Step S202: and a pupil image obtained by dividing the eye region. Specifically, the eye region image is further processed, and in the embodiment of the present invention, adaptive binarization processing is performed on the eye region. After binarization, the image only retains some pixel values in the pupil region and the eye region, and the rest of the pixel values are 0. And then processing the binarized image by using a region growing method, dividing the pupil region, and discarding pixels in a non-pupil region. In order to obtain more obvious image characteristics, the segmented pupils are subjected to morphological filtering, and pupil areas are obtained more accurately. Through the processing of the main steps, the extracted pupil position information and the extracted pupil image have more obvious characteristics, and the reliability of the subsequent model establishment is improved.
Specifically, in an embodiment, the step S102 specifically includes the following steps:
step S203: and carrying out image preprocessing on the images in the pupil information sample set. Specifically, in order to obtain more accurate data samples, image preprocessing is performed on the image data obtained in steps S201 to S202, so that irrelevant interference in the image is reduced. Wherein the pre-processing of the acquired image data includes, but is not limited to, image filtering, cropping, transforming, stitching and combining. The splicing combination comprises operations of splicing image areas, synthesizing image multiple channels and the like. And then, image data with obvious characteristics and smooth noise is obtained, and the reliability of the detection model established in the subsequent steps is improved.
Step S204: and performing data quality preprocessing on the data in the personnel information sample set. Specifically, in order to obtain a more accurate data sample, the preprocessing of the numerical data in the collected personnel information sample set mainly includes, but is not limited to, data cleaning, data integration, data specification, and data transformation. Therefore, unqualified data is removed, interference on the subsequent modeling step is avoided, and the reliability of the detection model established in the subsequent step is improved.
Step S205: and marking the pupil information sample set and the personnel information sample set according to the refraction data set, and constructing the marked samples into a detection training sample set. Specifically, the data in the staff information sample set and the image data in the pupil information sample set are matched according to the person to be collected, and then the refraction data, that is, the refraction information (including sphere lens, cylinder lens and axis position information, the obtaining scene of the information includes but is not limited to hospital, physical examination center, visual center, school physical examination, glasses shop, etc.) of the person to be collected is labeled correspondingly as the expected value of the training model of the matched data. And packaging the marked data to form a detection training sample set. The obtained detection training sample set has large data volume and delicate conditions, so that the refractive detection model is constructed in the step S103, the accuracy and the effectiveness of the adjusted model parameters are ensured, and the reliability of the refractive detection model is improved.
By executing the steps, the model training-based refraction detection method provided by the embodiment of the invention combines a large number of pupil images under different environmental factors with personal information of the collected person, so that sample data with conditions covering most conditions is obtained, the refraction detection model constructed by algorithms such as machine learning is complete as much as possible, the refraction detection can be effectively carried out on most persons to be detected, and the accuracy of the refraction detection is improved
As shown in fig. 3, the present embodiment further provides a model-based training refractive detection system applied to a refractive detection device, the system including:
the system comprises a data acquisition module 101 and a data acquisition module, wherein the data acquisition module acquires a pupil information sample set, a personnel information sample set and a refraction data set of a person to be acquired, the pupil information sample set at least comprises a pupil image and pupil position information, and the personnel information sample set at least comprises ambient illumination information and personnel station position information. For details, refer to the related description of step S101 in the above method embodiment, and no further description is provided here.
And the training sample module 102 is used for preprocessing data of the pupil information sample set and the personnel information sample set and constructing a detection training sample set according to the refraction data set. For details, refer to the related description of step S102 in the above method embodiment, and no further description is provided here.
And the detection module 103 trains the refraction detection model according to the model training sample set so as to perform refraction detection on the person to be detected through the trained refraction detection model. For details, refer to the related description of step S103 in the above method embodiment, and no further description is provided here.
Specifically, the refraction detection method based on model training provided by the embodiment of the invention further comprises the following steps:
step S206: a historical refractive data set is acquired. In particular, to further optimize the methods of the invention, the invention may provide a function of refraction prediction in addition to refraction detection. Acquiring historical refraction data of the acquired person to obtain the refraction information change trend of the acquired person in a period of time, wherein the acquisition scene of the data comprises but is not limited to hospitals, physical examination centers, visual centers, school physical examinations, spectacle shops and the like. In particular, the set of person information samples may further include: the name, class, sex, age, height, weight, BMI index, parent myopia, etc. of the person to be collected.
Step S207: and performing data quality preprocessing on the personnel information sample set, and constructing a prediction training sample set according to the historical refraction data set. Specifically, data quality detection is performed on data in the staff information sample set, so that a more accurate data sample is obtained, reliability of the prediction model is improved, then, the staff information sample set is labeled according to refractive information variation of the person to be collected in the historical refractive data within a period of time to construct a prediction training sample set, and the specific quality detection mode refers to step S204 and is not repeated.
Step S208: and training the vision prediction model according to the prediction training sample set so as to predict the vision of the person to be detected through the trained vision prediction model. The prediction model is trained by predicting historical vision data of a large number of collected persons in the training sample set, so that the reliability of sample data is improved, and the accuracy and the effectiveness of the prediction model are further improved. The detailed description of the prediction model construction method is the same as step S103, and is not repeated.
By executing the steps, the refraction detection method based on model training provided by the embodiment of the invention further builds a vision prediction model according to the trend of the vision of the collected person along with the change of time through historical vision data modeling, so that the method can realize future vision prediction besides refraction detection. Further improving the practicability of the method.
As shown in fig. 4, optionally, a model-training based refractive detection system further comprises:
a data acquisition module 206 acquires a sample set of staff information and a set of historical refractive data. For details, refer to the related description of step S206 in the above method embodiment, and no further description is provided herein.
And the training sample module 207 is used for carrying out data quality preprocessing on the personnel information sample set and constructing a prediction training sample set according to the historical refraction data set. For details, refer to the related description of step S207 in the above method embodiment, and no further description is provided here.
And the prediction module 208 trains the vision prediction model according to the prediction training sample set so as to predict the vision of the person to be detected through the trained vision prediction model. For details, refer to the related description of step S208 in the above method embodiment, and are not repeated herein.
The refractive detection system provided by the embodiment of the invention is used for executing the refractive detection method provided by the embodiment, the implementation manner and the principle are the same, and the details are referred to the related description of the embodiment of the method and are not repeated.
Through the cooperative cooperation of the above components, the model training-based refraction detection system provided by the embodiment of the invention combines the pupil images under a large number of different environmental factors with the personal information of the collected person, so as to obtain sample data with the situation covering most conditions, so that the refraction detection model constructed through algorithms such as machine learning is as complete as possible, the refraction detection can be effectively carried out on most persons to be detected, and the accuracy of the refraction detection is improved. And then, a vision prediction model is established according to the trend of the vision of the collected person changing along with time through historical refraction data modeling, so that the method can realize future vision prediction besides refraction detection. Further improving the practicability of the method.
FIG. 5 shows a refraction detection apparatus according to an embodiment of the invention, the apparatus comprising: the processor 901 and the memory 902 may be connected by a bus or other means, and fig. 5 illustrates the connection by the bus as an example.
The memory 902, which is a non-transitory computer-readable storage medium, may be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as program instructions/modules corresponding to the methods in the above-described method embodiments. The processor 901 executes various functional applications and data processing of the processor, i.e. implements the methods in the above-described method embodiments, by running non-transitory software programs, instructions and modules stored in the memory 902.
The memory 902 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created by the processor 901, and the like. Further, the memory 902 may include high speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory 902 may optionally include memory located remotely from the processor 901, which may be connected to the processor 901 via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
One or more modules are stored in the memory 902, which when executed by the processor 901 performs the methods in the above-described method embodiments.
The specific details of the refraction detecting device can be understood by referring to the corresponding related description and effects in the above method embodiments, and are not described herein again.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, and the implemented program can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic Disk, an optical Disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a Flash Memory (Flash Memory), a Hard Disk Drive (Hard Disk Drive, abbreviated as HDD), or a Solid State Drive (SSD); the storage medium may also comprise a combination of memories of the kind described above.
Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations fall within the scope defined by the appended claims.
Claims (10)
1. A method for refractive detection based on model training, the method comprising:
acquiring a pupil information sample set, a personnel information sample set and a refraction data set of a person to be acquired, wherein the pupil information sample set at least comprises a pupil image and pupil position information, and the personnel information sample set at least comprises environmental illumination information and personnel station information;
carrying out data preprocessing on the pupil information sample set and the personnel information sample set, and constructing a detection training sample set according to the refraction data set;
and training the refraction detection model according to the detection training sample set so as to perform refraction detection on the person to be detected through the trained refraction detection model.
2. The method of claim 1, wherein the acquiring of the pupil information sample set at least comprises a pupil image and pupil position information, comprising:
positioning the eye region of the face image of the person to be collected to obtain the pupil position information;
and the pupil image is obtained by dividing the eye region.
3. The method according to claim 1, wherein the data preprocessing the pupil information sample set and the person information sample set to construct a detection training sample set comprises:
carrying out image preprocessing on the images in the pupil information sample set;
performing data quality preprocessing on the data in the personnel information sample set;
and labeling the pupil information sample set and the personnel information sample set according to the refraction data set, and constructing the labeled samples into the detection training sample set.
4. The method of claim 3, wherein the labeling the pupil information sample set and the people information sample set from the refraction data set comprises:
the refractive data set includes refractive values including sphere, cylinder, and axis information.
5. The method of claim 1, wherein the sample set of personal information further comprises at least personal information of the subject.
6. A model-training based refractive detection system, the system comprising:
the data acquisition module is used for acquiring a pupil information sample set, a personnel information sample set and a refraction data set of a person to be acquired, wherein the pupil information sample set at least comprises a pupil image and pupil position information, and the personnel information sample set at least comprises environmental illumination information and personnel station position information;
the training sample module is used for carrying out data preprocessing on the pupil information sample set and the personnel information sample set and constructing a detection training sample set according to the refraction data set;
and the detection module trains the refraction detection model according to the model training sample set so as to perform refraction detection on the person to be detected through the trained refraction detection model.
7. A method for model-based training for refractive detection, the method further comprising:
acquiring the personnel information sample set and the historical refraction data set;
performing data quality pretreatment on the personnel information sample set, and constructing a prediction training sample set according to the historical refraction data set;
and training the vision prediction model according to the prediction training sample set so as to predict the vision of the person to be detected through the trained vision prediction model.
8. A model-training based refractive detection system, the system comprising:
the data acquisition module is used for acquiring the personnel information sample set and the historical refraction data set;
the training sample module is used for preprocessing the data quality of the personnel information sample set and constructing a prediction training sample set according to the historical refraction data set;
and the prediction module is used for training the vision prediction model according to the prediction training sample set so as to predict the vision of the person to be detected through the trained vision prediction model.
9. A refractive detection apparatus, comprising:
a memory and a processor, communicatively coupled to each other, the memory having stored therein computer instructions, the processor executing the computer instructions to perform the method of any of claims 1-5 and 7.
10. A computer-readable storage medium having stored thereon computer instructions for causing a computer to thereby perform the method of any one of claims 1-5 and 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110563271.8A CN115375611A (en) | 2021-05-21 | 2021-05-21 | Model training-based refraction detection method and detection system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110563271.8A CN115375611A (en) | 2021-05-21 | 2021-05-21 | Model training-based refraction detection method and detection system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115375611A true CN115375611A (en) | 2022-11-22 |
Family
ID=84058916
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110563271.8A Pending CN115375611A (en) | 2021-05-21 | 2021-05-21 | Model training-based refraction detection method and detection system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115375611A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116028870A (en) * | 2023-03-29 | 2023-04-28 | 京东方艺云(苏州)科技有限公司 | Data detection method and device, electronic equipment and storage medium |
-
2021
- 2021-05-21 CN CN202110563271.8A patent/CN115375611A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116028870A (en) * | 2023-03-29 | 2023-04-28 | 京东方艺云(苏州)科技有限公司 | Data detection method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9779492B1 (en) | Retinal image quality assessment, error identification and automatic quality correction | |
| CN109684915B (en) | Pupil tracking image processing method | |
| WO2021135557A1 (en) | Artificial intelligence multi-mode imaging analysis apparatus | |
| WO2017020045A1 (en) | System and methods for malarial retinopathy screening | |
| CN110575132A (en) | Method for calculating degree of strabismus based on eccentric photography | |
| CN110575134A (en) | method for calculating myopia degree based on eccentric photography | |
| CN115375611A (en) | Model training-based refraction detection method and detection system | |
| Abdani et al. | Iris segmentation method of pterygium anterior segment photographed image | |
| CN117338234A (en) | Diopter and vision joint detection method | |
| CN110503636B (en) | Parameter adjustment method, focus prediction method, parameter adjustment device and electronic equipment | |
| Kumar et al. | Detection of glaucoma using image processing techniques: a critique | |
| Nair et al. | Multi-labelled ocular disease diagnosis enforcing transfer learning | |
| CN115456974A (en) | Strabismus detection system, method, equipment and medium based on face key points | |
| Hagos et al. | Automated smartphone based system for diagnosis of diabetic retinopathy | |
| KR102282334B1 (en) | Method for optic disc classification | |
| Haloi | Towards ophthalmologist level accurate deep learning system for OCT screening and diagnosis | |
| Alam et al. | Benchmarking Deep Learning Frameworks for Automated Diagnosis of Ocular Toxoplasmosis: A Comprehensive Approach to Classification and Segmentation | |
| CN109447995B (en) | Anterior segment image segmentation method and related device thereof | |
| Caya et al. | Development of Pupil Diameter Determination using Tiny-YOLO Algorithm | |
| Shanthakumari et al. | Glaucoma Detection using Fundus Images using Deep Learning | |
| CN113116292B (en) | Eye position measurement method, device, terminal and equipment based on eye appearance image | |
| US20220151482A1 (en) | Biometric ocular measurements using deep learning | |
| Viraktamath et al. | Detection of Diabetic Maculopathy | |
| Khan et al. | Cost-Effective early warning solution for Anisocoria Eye-Disease through Optical Sensing and Machine Learning: A Preliminary Analysis | |
| US20220180509A1 (en) | Diagnostic tool for eye disease detection |
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 |