CN-121971280-A - Multi-dimensional visual training feedback system and method for improving binocular fusion vision and stereoscopic vision functions

CN121971280ACN 121971280 ACN121971280 ACN 121971280ACN-121971280-A

Abstract

The invention belongs to the technical field of vision training, and particularly relates to a multidimensional vision training feedback method for improving binocular fusion vision and stereoscopic vision functions, which comprises the following steps of detecting a medical baseline, generating an initial capability portrait through vision function evaluation and cognition capability test, matching training difficulty parameters, and generating a personalized scheme; the training method comprises the steps of performing multi-mode game training, training types including game training, animation scene training and offline family task training, and dynamically adjusting training difficulty, guiding a teacher remotely one by one, dynamically adjusting a training scheme according to a change report by tracking training progress in real time, and performing multi-dimensional effect evaluation and self-optimizing a difficulty prediction algorithm by the system. The invention solves the problems of boring and passive participation in training, scheme rigidification and execution deviation, single stimulation and cognitive load, and limited and continuous scene interruption.

Inventors

WANG JIANGBO
WANG JIAO
QIU XIAOJUAN
WANG CHANGZAI

Assignees

杭州深康明视科技有限公司

Dates

Publication Date: 20260505
Application Date: 20260206

Claims (10)

1. A multi-dimensional visual training feedback method for improving binocular fusion vision and stereoscopic vision functions is characterized by comprising the following specific steps: Performing medical baseline detection, generating an initial capability portrait through visual function evaluation and cognitive ability test, matching training difficulty parameters, and generating a personalized scheme; Performing multi-mode game training, wherein the training types comprise gambling training, animation scene training and offline family task training, and dynamically adjusting training difficulty; the teacher carries out one-to-one guidance remotely, and dynamically adjusts the training scheme according to the change report by tracking the training progress in real time; The system performs multidimensional effect evaluation and self-optimizes a difficulty prediction algorithm.
2. The multi-dimensional visual training feedback method for improving binocular fusion vision and stereoscopic vision functions according to claim 1, wherein the gambling training comprises dynamic tracking and saccadic training, red-blue/polarization fusion training and random point stereogram training.
3. The multi-dimensional visual training feedback method for improving binocular fusion vision and stereoscopic vision functions according to claim 1, wherein the specific steps of the game training are as follows: Based on a hierarchical management principle of age and visual severity, and combining game resources, adopting a game engine to construct an interactive task and a jaywalking mechanism; Integrating a touch screen, eye tracking, gesture recognition and dynamic scene simulation, responding to the interactive operation of the child game, and adaptively adjusting the interactive mode according to the vision of the child and the game tracking condition; executing a core vision training task, dynamically adjusting training difficulty according to user data, amblyopia type and age, and selecting game content according to individuality of children; Storing and processing user training data, calculating performance indexes in real time, triggering difficulty adjustment, and assisting in optimizing a personalized training scheme; and synchronously testing the influence of different interface versions on the interest of children.
4. The multi-dimensional visual training feedback method for improving binocular fusion vision and stereoscopic vision functions according to claim 3, wherein the construction process of the interactive task and the jaywalking mechanism further comprises automatically generating a new checkpoint or task variant based on templates and randomization technology, and the algorithm can adjust color saturation or target moving speed to create new challenges.
5. The multi-dimensional visual training feedback method for improving binocular fusion vision and stereoscopic vision functions according to claim 1, wherein the specific steps of the animation scene training are as follows: Generating animation content, adopting a technology of converting a static microscope into a dynamic video, implanting the animation into a visual training element, adjusting material reflectivity and illumination parameters, simulating the light and shadow effect of a real environment, analyzing a amblyopia training target based on an NLP model, generating a scenario script, and integrating a visual task into a story line; Stimulating the development of cone cells by adopting a color therapy, training the following and glancing capacities of eyeballs by adopting a dynamic optotype, training the functions of binocular vision by adopting a red-blue split vision or polarized light technology, promoting the development of fusion vision and stereoscopic vision by adopting a 3D parallax scene and parallax change technology, and training the peripheral area of retina by adopting a wide-angle lens and dynamic edge light shadow technology; designing a scenario interaction task, detecting the operation accuracy of a user in real time, triggering vibration, excitation sound effect and positive feedback of virtual rewards when the task is completed, and dynamically adjusting the scenario rhythm according to eye movement tracking data; analyzing preference and capability, pushing adaptive animation, analyzing performance data every 30 seconds, automatically adjusting animation task parameters, generating vision progress curve, and synchronizing to a parent terminal APP and a teacher system.
6. The multi-dimensional visual training feedback method for improving binocular fusion vision and stereoscopic vision functions of claim 5, wherein the generating process of the animation content further comprises the steps of establishing a database to store animation resources, grading and labeling according to age, amblyopia type and severity, synchronously generating animation variants, and avoiding content repetition.
7. The multi-dimensional visual training feedback method for improving binocular fusion vision and stereoscopic vision functions according to claim 1, wherein the specific steps of the teacher remote one-to-one guidance are as follows: performing multidimensional evaluation and baseline establishment, and assisting in customizing a personalized training plan; In addition, parents upload auxiliary training videos through the APP, when AI identifies continuous 3 times of task failure or progress stagnation, the system pushes intervention prompts to the guide operators, and a visual report is generated; The system AI automatically promotes and degrades the data by analyzing the data once every 30 seconds, guides the parents to adjust through the active intervention of video consultation, revises the training key point by combining Zhou Dushu report, and pushes the new edition plan to the parents.
8. The method for improving binocular fusion vision and stereoscopic vision according to claim 7, wherein the specific process of establishing the multi-dimensional evaluation and the base line comprises the steps of calling an initial vision test report of a child, recommending initial training combination by a system in combination with a first game task performance, and adjusting parameters by a teacher on a digital platform to generate a custom plan comprising frequency and family auxiliary training.
9. A multi-dimensional visual training feedback system for improving binocular fusion vision and stereoscopic vision functions and executing the multi-dimensional visual training feedback method according to any one of claims 1-8, wherein the visual training feedback system comprises a user interaction layer, a data processing layer, an AI engine layer and a service support layer; The user interaction layer is used for multi-mode perception and feedback, and is provided with a VR/AR immersion training terminal and a multi-sense feedback device, wherein the VR/AR immersion training terminal generates a dynamic stereoscopic visual target by being provided with an eye movement tracking module and a 3D display technology; The data processing layer is used for analyzing and personalized management in real time, integrating a multi-source data fusion center and a personalized scheme generation engine, wherein the multi-source data fusion center is used for integrating ophthalmic detection data, training logs and physiological indexes, the personalized scheme generation engine is used for matching training combination based on amblyopia type and age, and an AI model optimizes parameters every 30 seconds; The AI engine layer is used for intelligent training and decision assistance, and is provided with a multi-mode vision-language big model and a vision function training module, wherein the multi-mode vision-language big model is used for analyzing OCT/fundus imaging, identifying the cell liveness of a macular region, generating a training report and responding to medical questions and answers, federal learning integrates global doctor feedback, and the vision function training module comprises simultaneous vision training, fusion vision training and stereoscopic vision reconstruction training; The service support layer is used for collaborative management and compliance guarantee, a remote teacher collaboration platform and a family-medical linkage system are built, wherein the remote teacher collaboration platform supports video guidance, a whiteboard is shared to mark eyeball movement tracks, training parameters are dynamically adjusted, and a parent end APP is developed in the family-medical linkage system to realize pushing family auxiliary training.
10. An electronic device is characterized by comprising at least one processor; and a memory communicatively coupled to the at least one processor; Wherein the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the multi-dimensional visual training feedback method of any one of claims 1 to 8.

Description

Multi-dimensional visual training feedback system and method for improving binocular fusion vision and stereoscopic vision functions Technical Field The invention belongs to the technical field of vision training, and particularly relates to a multidimensional vision training feedback system and method for improving binocular fusion vision and stereoscopic vision functions. Background The multi-dimensional visual training is a comprehensive visual function intervention method, improves coordination and efficiency of a visual system through multi-angle and multi-level targeted training, enhances focusing capacity, stereoscopic vision and dynamic visual response of eyes through adjusting training (such as far and near focus switching), fusion training (binocular coordination) and eyeball movement training (tracking and saccadic), relaxes eye muscles through dynamic adjusting training (such as reverse shooting and crystal manipulation) aiming at ciliary muscle spasm caused by long-term short-distance eye use, delays myopia development, improves visual fatigue symptoms, is applied to amblyopia (to improve visual acuity), strabismus operation (to reconstruct binocular vision function) and the like, and combines disinhibition training and nerve plasticity stimulation. Problems of the prior art: the problems of boring training and passive participation in the traditional vision training are that repeated mechanical training leads to the contradiction of children, and the compliance is low (the falling rate is more than 40 percent); The scheme stiffness and execution deviation problem is that the traditional vision training is limited in that the family training lacks professional supervision, the adjustment lag (such as monthly review) and the individuation is insufficient; the problem of single stimulus and cognitive load is that the traditional visual training is limited in that the static visual stimulus is single in dimension and is difficult to maintain attention, and the high-strength and fast-paced training is easy to lead children to have training resistant psychology; The problems of limited scene and continuous interruption are that the traditional visual training is limited to relying on hospital instruments, and the family training is difficult to link. Disclosure of Invention The invention aims to provide a multidimensional visual training feedback system and a multidimensional visual training feedback method for improving binocular fusion vision and stereoscopic vision functions, which can solve the problems of boring and passive participation in training, scheme rigidification and execution deviation, single stimulation and cognitive load, and scene limitation and continuous interruption. The technical scheme adopted by the invention is as follows: A multi-dimensional visual training feedback method for improving binocular fusion vision and stereoscopic vision functions comprises the following specific steps: Performing medical baseline detection, generating an initial capability portrait through visual function evaluation and cognitive ability test, matching training difficulty parameters, and generating a personalized scheme; Performing multi-mode game training, wherein the training types comprise gambling training, animation scene training and offline family task training, and dynamically adjusting training difficulty; the teacher carries out one-to-one guidance remotely, and dynamically adjusts the training scheme according to the change report by tracking the training progress in real time; The system performs multidimensional effect evaluation and self-optimizes a difficulty prediction algorithm. The gambling training comprises dynamic tracking and glancing training, red-blue/polarization fusion training and random point stereogram training. The specific steps of the game training are as follows: Based on a hierarchical management principle of age and visual severity, and combining game resources, adopting a game engine to construct an interactive task and a jaywalking mechanism; Integrating a touch screen, eye tracking, gesture recognition and dynamic scene simulation, responding to the interactive operation of the child game, and adaptively adjusting the interactive mode according to the vision of the child and the game tracking condition; executing a core vision training task, dynamically adjusting training difficulty according to user data, amblyopia type and age, and selecting game content according to individuality of children; Storing and processing user training data, calculating performance indexes in real time, triggering difficulty adjustment, and assisting in optimizing a personalized training scheme; and synchronously testing the influence of different interface versions on the interest of children. The construction process of the interactive task and the intrusion mechanism further comprises the steps of automatically generating a new checkpoint or task variant based on templates