A virtual reality-based personalized visual training system for visual field loss

Abstract

The application discloses a virtual reality-based visual field defect personalized visual training system and relates to the field of visual training. When visual training is performed, a trainee performs visual training according to a corresponding generated visual training video on a display device through a head-mounted VR glasses. In the process of performing visual training according to the generated visual training video, a cloud data analysis platform performs gravity compensation and motion energy integration according to head acceleration data. When the motion energy integration is greater than an energy threshold, the display device is controlled to pause the visual training video for a set time. According to reaction time and success rate of different sites determined according to handle button event data, a scene generation and video processing module is controlled to dynamically adjust the visual training video. The application can realize personalized customization and dynamic adjustment of training content, and has the characteristics of simple operation, low training cost and good training effect.

Description

Technical Field

[0001] This application relates to the field of visual training, and in particular to a personalized visual training system for visual field defects based on virtual reality. Background Technology

[0002] Glaucoma is not only associated with the loss of retinal ganglion cells, but also with the degeneration of cortical and subcortical brain structures related to eye movements. Eye movements can be categorized into saccades, fixations, and smooth tracking. Saccades refer to the rapid guidance of the gaze to a target of interest, while fixations allow for highly sensitive observation of the target to extract visual information. The fixation point can be static (for stationary targets) or smooth tracking (for moving targets). Numerous studies have shown that glaucoma patients exhibit abnormal eye movement patterns. Research indicates that glaucoma patients have longer saccade latency, which increases with the severity of glaucoma; reduced saccade amplitude and speed; and the presence of uncontrollable reflexive saccades. Fixation stability is also reduced. When relying solely on eye movements for visual target searching, glaucoma patients tend to perform saccades with lower frequency, lower speed, and longer reaction times.

[0003] Fixation reflects an individual's ability to spontaneously maintain gaze on a target, thus reflecting cognitive control. Sagging enables rapid searching of the visual field and selection of stimulus information; the speed and accuracy of saccades reflect the function of the frontal lobe and brainstem. Following movements are controlled by a different set of neural mechanisms than saccades, with key brain regions including the midtemporal visual cortex and the medial superior temporal cortex. The cerebellum also plays a crucial role in maintaining the accuracy and fluency of these movements. Fixation training aims to improve the stability of fixation function and attention. Sagging training can improve saccade efficiency, reduce latency, increase accuracy, and induce spontaneous saccades during head movements to compensate for insufficient vestibular-ocular reflexes. Good following function ensures that moving targets in space are continuously and stably imaged in the fovea during eye movements.

[0004] Visual training systems for glaucoma have also emerged, and existing visual training systems mainly reflect the following aspects:

[0005] 1. Computer-based visual perception training: This method uses Gabor icons on contrast thresholds for stimulation, aiming to improve visual function and quality of life in glaucoma patients. The training group received more challenging contrast threshold training, while the control group received the easiest contrast training. Patients were required to practice for 30 minutes every night before bed and were followed up continuously.

[0006] 2. RevitalVision Visual Training: This is a therapeutic technique based on visual stimulation and promoting neural connectivity responses. It improves visual quality by promoting neuronal connections in the visual cortex through online computer-based visual training, thereby enhancing the processing capacity of neurons in the visual cortex. Specific steps include pre-training assessment, customizing a training plan based on individual circumstances, conducting training, and follow-up examinations. Training content includes single-image visual training, clear-image visual training, triple-image visual training, and moving-image visual training.

[0007] 3. Virtual Reality Binocular Vision Push-Pull Model Training: This method uses virtual reality technology for visual training, aiming to improve stereoscopic vision in glaucoma patients. Patients undergo visual training twice daily for 20 minutes each time, with follow-up assessments conducted one week and one month after training.

[0008] However, existing visual training systems have the following drawbacks:

[0009] 1. Adaptation issues: Some patients may have difficulty adapting to compressed images, which can cause distortion of spatial and distance perception. Patients need to adapt to this limitation of the system to complete daily tasks.

[0010] 2. Uncertainty of effects: Although some studies have shown that visual training can improve visual function in glaucoma patients, the long-term stability and practical application value of these effects still need to be determined by large-sample studies.

[0011] 3. Technical limitations: Traditional visual aids that expand the field of vision have drawbacks such as fixed magnification, inability to enhance image contrast, and distortion of spatial and distance perception. These limitations also affect the effectiveness of visual training.

[0012] 4. Training continuity and compliance: Visual training requires patients to persist in it for a long time, but patients may find it difficult to continue training for various reasons (such as the boredom of training, time cost, etc.).

[0013] 5. Equipment cost and accessibility: Some visual training methods may require expensive equipment, which may limit their application in resource-constrained areas.

[0014] Based on the above problems, there is an urgent need to provide a new visual training system that can realize personalized customization of training content and has the characteristics of simple operation, low training cost and good training effect. Summary of the Invention

[0015] The purpose of this application is to provide a personalized visual training system for visual field defects based on virtual reality, which can realize personalized customization of training content and has the characteristics of simple operation, low training cost and good training effect.

[0016] To achieve the above objectives, this application provides the following solution:

[0017] This application provides a virtual reality-based personalized visual training system for visual field defects, which includes: a head-mounted VR glasses, a display device, a scene generation and video processing module, a head motion calculation module, a controller interaction module, and a cloud data analysis platform.

[0018] The scene generation and video processing module is used to determine the visual field defect area based on the visual field detection results of the trainee; and to generate a visual training video based on the Unity3D platform according to the visual field defect area; the visual training video includes 3D training videos of different visual training tasks and left and right modes.

[0019] The scene generation and video processing module is mounted on a cloud server and sends the generated visual training video to the visual training APP installed on the display device.

[0020] The VR headset is worn over the eyes of the person being trained.

[0021] The head motion calculation module is used to acquire head acceleration data through the mobile phone accelerometer; and to perform gravity compensation and motion energy integration based on the head acceleration data.

[0022] The handle interaction module is used to provide feedback via haptic feedback buttons;

[0023] The cloud-based data analysis platform is connected to the head motion calculation module, the controller interaction module, the scene generation and video processing module, and the display device, respectively; and is used to store head acceleration data and controller button event data, and generate a comprehensive evaluation report.

[0024] During visual training, the trainee uses VR headsets to perform visual training based on the corresponding visual training videos generated on the display device. During the visual training process, the cloud data analysis platform performs gravity compensation and motion energy integration based on head acceleration data. When the motion energy integration exceeds the energy threshold, the display device is controlled to pause the visual training video for a set time. Based on the reaction time and success rate of different points determined by the controller button event data, the scene generation and video processing modules are controlled to dynamically adjust the visual training video.

[0025] Optionally, the sampling rate of the haptic feedback button is ≥100Hz.

[0026] Optionally, the controller interaction module uses Bluetooth for interaction.

[0027] Optionally, the display device is a mobile phone.

[0028] Optionally, the scene generation and video processing module uses Adobe Premiere Pro for visual training video processing.

[0029] Optionally, the 3D training videos include: visual field stimulation instructions, right eye visual field stimulation, rest after removing the eye mask, left eye visual field stimulation, rest after removing the eye mask and closing the eyes, eye movement training instructions, saccade fixation, visual tracking, and visual matching.

[0030] Optionally, the scene generation and video processing module employs the SpawnAndDestroyBall function; the SpawnAndDestroyBall function is used to cyclically generate personalized stimuli for the visual stimulation portion in the visual training video.

[0031] Optionally, the scene generation and video processing module uses Visual Studio 2022 as the compiler for C# scripts.

[0032] Optionally, the process of dynamically adjusting the visual training video by the cloud-based data analysis platform's control scene generation and video processing module is as follows:

[0033] When visual stimulation occurs, if the reaction time at the current site is greater than the reaction time threshold or the success rate is less than 50%, the duration and probability of the stimulus ball appearing in the visual training video are increased to 20% of the original visual training video, and the size of the stimulus ball is increased to a 2.5° angle of view. After three consecutive successful attempts, the original visual training video is restored.

[0034] During eye-tracking training, when the reaction time of the ball moving from one point to another along the saccade fixation path is greater than the reaction time threshold or the success rate is less than 50%, the probability of the saccade fixation path appearing is increased to 20% of the original visual training video, stimulating the ball size to increase to a 2.5° angle of view. After three consecutive successful attempts, the original visual training video is restored.

[0035] According to the specific embodiments provided in this application, this application has the following technical effects:

[0036] This application provides a personalized visual training system for visual field defects based on virtual reality. The system uses a scene generation and video processing module to determine the visual field defect area based on the visual field detection results of the trainee. Based on the visual field defect area, a visual training video is generated using the Unity3D platform. This enables the development of personalized virtual reality training content using the Unity3D platform, providing a low-cost, high-efficiency immersive visual rehabilitation solution for patients with visual field defects. The system achieves personalized customization of training content and demonstrates significant clinical application value. Simultaneously, the cloud-based data analysis platform of this application controls the scene generation and video processing module to dynamically adjust the visual training video based on head acceleration data and controller button event data, allowing for dynamic adjustment of training parameters according to real-time performance. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This is a schematic diagram of a personalized visual training system for visual impairment based on virtual reality in one embodiment of this application;

[0039] Figure 2 This is a screenshot illustration of a portion of the video showing visual stimulation.

[0040] Figure 3 A screenshot illustration of the part of the video being scanned and focused ( Figure 3 Part (a) is a screenshot of the saccade fixation stimulus ball at the center of the visual field. Figure 3 (b) is a screenshot of the peripheral visual field of the saccade fixation stimulation ball.

[0041] Figure 4 This is a screenshot illustration of the visual matching part of the video. Detailed Implementation

[0042] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0043] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0044] In one exemplary embodiment, such as Figure 1 As shown, a personalized visual training system for visual field defects based on virtual reality is provided. The system includes: a head-mounted VR glasses, a display device, a scene generation and video processing module, a head motion calculation module, a controller interaction module, and a cloud data analysis platform.

[0045] The scene generation and video processing module is used to determine the visual field defect area based on the visual field detection results of the trainee; and based on the visual field defect area, it generates visual training videos on the Unity3D platform. The visual training videos include different visual training tasks and 3D training videos in left-right mode. The left-right mode 3D training videos can avoid interference (compensation effect) from the contralateral eye on the training eye, improve the eye's accommodation ability and visual spatial awareness, and reduce eye fatigue and visual pressure. The left-right mode 3D training videos are recorded separately, and a single eye patch is used to cover the non-training eye during training to ensure that the training of each eye is conducted independently. The scene generation and video processing module is mounted on a cloud server and sends the generated visual training videos to the visual training APP installed on the display device.

[0046] The process for determining the 3D training video in left-right mode is as follows:

[0047] 1. In the Unity editor's Hierarchy panel, select GameObject > Camera, create two Camera objects to simulate the perspective of the left and right eyes respectively, and set their interpupillary distance to 60mm to simulate the normal distance of the human eye.

[0048] 2. Combine the left and right eye images into a single image by creating a StereoRender Texture with a resolution of 2560x1280. The left half displays the left eye's view, and the right half displays the right eye's view. StereoRenderTexture is used for stereo rendering. It allows the left and right eye views to be rendered into the same texture separately, thus achieving the stereoscopic visual effects commonly found in VR or AR applications.

[0049] 3. Assign the Stereo RenderTexture to the Target Texture property of both Cameras, and the left and right eye views will be rendered onto that Texture respectively. Target Texture is a property in Unity used to specify the rendering target of a camera;

[0050] 4. Click “Start Recording” to begin recording. Unity will automatically start capturing and synthesizing video clips from the left and right eye perspectives, recording visual training videos for the two modules of Part 2 for each user.

[0051] The scene generation and video processing module is mounted on the display device;

[0052] The VR headset is worn over the eyes of the person being trained;

[0053] During visual training, the trainee uses VR glasses to perform visual training based on the corresponding visual training video generated on the display device. This application uses virtual reality technology to accurately locate and stimulate the patient's ARV area to activate residual visual function and help visually impaired patients improve their visual perception ability.

[0054] The head motion calculation module is used to acquire head acceleration data through the mobile phone accelerometer; and to perform gravity compensation and motion energy integration based on the head acceleration data.

[0055] The controller interaction module is used to provide feedback through haptic feedback buttons; the controller interaction module communicates via Bluetooth and has a sampling rate ≥100Hz;

[0056] The cloud-based data analysis platform is connected to the head motion calculation module, the controller interaction module, the scene generation and video processing module, and the display device, respectively; and is used to store head acceleration data and controller button event data, and generate a comprehensive evaluation report.

[0057] During visual training, the trainee uses VR headsets to perform visual training based on the corresponding visual training video generated on the display device. During the visual training process, the cloud data analysis platform performs gravity compensation and motion energy integral E based on head acceleration data. When the motion energy integral E is greater than the energy threshold, the display device is controlled to pause the visual training video for a set time (10 seconds). Based on the reaction time and success rate of different points determined by the controller button event data, the scene generation and video processing modules are controlled to dynamically adjust the visual training video.

[0058] The handle interaction module achieves real-time evaluation in the following ways: including visual field stimulation and saccade fixation during eye movement training: when the ball is seen, pressing the button automatically calculates the reaction time and success rate at different points in the cloud (calculated by combining the performance of the previous 2 days, i.e., the performance of the 4 complete training sessions), and can dynamically adjust the probability of different stimulation points and the stimulation duration according to the subject's performance; when the ball is seen, pressing the button transmits the data to the cloud to calculate the reaction time and success rate to evaluate the subject's eye movement function.

[0059] In a specific embodiment, the dynamic adjustment process is as follows:

[0060] During the initial training phase (the first 5 days of training), the training parameters remained unchanged.

[0061] Subsequently, when visual stimulation occurs, if the reaction time at the current location is greater than the reaction time threshold or the success rate is less than 50%, the duration and probability of the stimulus ball appearing in the visual training video are increased to 20% of the original visual training video, and the size of the stimulus ball is increased to a 2.5° angle of view. After three consecutive successful attempts, the original visual training video is restored.

[0062] During eye-tracking training, when the reaction time of the ball moving from one point to another along the saccade fixation path is greater than the reaction time threshold or the success rate is less than 50%, the probability of the saccade fixation path appearing is increased to 20% of the original visual training video, and the ball size is increased to a 2.5° field of view. After three consecutive successful attempts, the original visual training video is restored.

[0063] In one specific embodiment, the head-mounted VR glasses are the Qianhuan Magic Mirror VR phone box. The Qianhuan Magic Mirror VR phone box features a lightweight design and high-precision aspherical optical lenses. It is made of a new type of ABS material in a single mold, weighing only 231g, making it lightweight, comfortable, and pressure-free to wear. It employs a unique optical system with aspherical lenses, providing anti-radiation and anti-blue light protection, effectively reducing visual distortion and edge blurring, and offering a clear and natural stereoscopic visual effect.

[0064] In another embodiment, a mobile phone is used as a display device, and an immersive virtual reality training environment is constructed by combining the Qianhuan Magic Mirror VR mobile phone box and the mobile phone. This application can achieve an immersive 3D virtual reality experience with relatively low hardware costs.

[0065] The Qianhuan Magic Mirror VR mobile phone box includes: an interpupillary distance (IPD) adjustment button, an object distance (OD) adjustment button, a phone shield, a headband, and optical lenses. By rotating the IPD and ODD buttons forward and backward, users can simultaneously adjust both IPD and focal length: the IPD adjustment range is 58.5mm to 70.5mm, and the focal length adjustment is suitable for users with myopia up to 600 degrees, adapting to the visual needs of different users and enhancing the device's versatility and practicality. Furthermore, the enclosed structure design effectively shields against ambient light interference, further enhancing visual immersion; while the lightweight body, optimized wearing structure, and use of high-precision aspherical lenses significantly improve user comfort. Leveraging the high-resolution display of the mobile phone, this application enables the development of personalized virtual reality training content, providing a low-cost, high-efficiency immersive visual rehabilitation solution for patients with visual field defects, demonstrating significant clinical application value.

[0066] The visual training scenario in this application was designed using Unity 3D programming, and two experimental scenarios were developed: visual stimulation and eye-tracking training. Unity is a cross-platform interactive content creation tool developed by Unity Technologies. The Unity game engine supports development on various desktop, mobile, console, and virtual reality platforms, and is popular in many fields due to its excellent interactivity, adaptability, practicality, and technological innovation. The engine provides the main scripting API for C#, and scripts can be compiled using UnityScript or the more general Visual Studio. This application uses Visual Studio 2022 as the C# script compiler.

[0067] To ensure efficient integration and smooth transition of the visual training video content, this application uses Adobe Premiere Pro (PR) for post-production video processing. As a leading non-linear video editing software, PR offers significant advantages in multi-track editing, special effects applications, and format compatibility. In this application, PR is primarily used for precise editing and integration of the two core training videos: "Visual Field Stimulation" and "Eye Tracking Training," ensuring the logical coherence of the training content. Furthermore, by embedding explanatory subtitles, suggestive animations, and rest time guidance into the videos, the patient's understanding of the training tasks and ease of operation are further enhanced.

[0068] The visual training videos in this application are divided into two core modules: visual field stimulation and eye movement training. The visual field stimulation module designs a personalized stimulation plan for each enrolled patient based on the location and extent of the visual field defect shown in their visual field results. This aims to precisely activate residual visual field function and improve visual perception. The eye movement training module includes three stages: saccade fixation, visual tracking, and visual matching. This strengthens eye movement coordination, improves dynamic target tracking ability, and enhances visual information integration efficiency, thereby improving the user's eye movement function and reaction speed. Through the synergistic training of these two modules, while improving patients' visual acuity, it significantly enhances eye movement function, providing an efficient and scientific rehabilitation solution for patients with visual field defects.

[0069] The visual training video includes a total 3D training video with a duration of 19 minutes and 42 seconds, consisting of the following parts in sequence: visual field stimulation explanation, right eye visual field stimulation (5 minutes and 25 seconds), rest after removing the eye mask, left eye visual field stimulation (5 minutes and 25 seconds), rest after removing the eye mask and closing the eyes (1 minute), eye movement training explanation, saccade fixation (1 minute and 50 seconds), visual tracking (1 minute and 50 seconds), and visual matching (1 minute and 50 seconds).

[0070] Specifically, such as Figure 2As shown, in the visual training video's visual stimulation section, the background color is set to gray (RGB[50,50,50]) to create a sharp contrast with the stimulation. A green cross (RGB[50,200,50]) is displayed in the center of the visual field, the size of a V-shaped visual target, corresponding to a visual angle of approximately 1.72°. The user must keep their eyes fixed on the green cross. The generation location of the stimulation ball is determined based on the user's visual field detection results; the stimulation ball randomly appears in a specific area within a plane 5 meters away; the stimulation ball is a white ball (RGB[255,255,255]), its size equal to the V-shaped visual target, corresponding to a visual angle of approximately 1.72°; each stimulation ball uses the SpawnAndDestroyBall function, continuously displayed on the screen for 2 seconds before disappearing, and then generating the next ball after a 0.2-second interval. The entire process is repeated cyclically to ensure continuous stimulation of the user's target area and accurate activation of residual visual function.

[0071] Specifically, such as Figure 3 As shown, in the saccade fixation section of the visual training video, the stimulus ball is a red (RGB[255,0,0]) ball, the size of which is a V-shaped optotype, corresponding to a visual angle of approximately 1.72°. The stimulus ball first appears at the center of a plane 5 meters away from the user, is displayed for 2 seconds, and then disappears. After each disappearance, the ball randomly selects a new position within a 30° range (up, down, left, and right) of the visual field angle and re-displays at that position for 2 seconds. This process is repeated; the user needs to follow the movement trajectory of the stimulus ball with their eyes to complete the fixation and saccade tasks. The user's eye movements should follow the dynamics of the stimulus ball to maintain stable fixation and saccades. This part trains the user's eye movement control ability, while promoting the visual system's ability to quickly identify and respond to targets, further improving the coordination of fixation and saccades.

[0072] Specifically, in the visual tracking section of the visual training video, the stimulus ball is a red ball (RGB[255,0,0]) with a field of view of approximately 1.72°. The stimulus ball will move sequentially along a zigzag path (horizontal, vertical, 45° diagonal, and -45° diagonal) within a plane 5 meters of the user, completing a full zigzag path. After moving from the starting point to the endpoint in the set direction, the stimulus ball will return to the starting point along the same path, completing one round trip. After each round trip, the stimulus ball is destroyed, and a new stimulus ball is generated after a short interval, and the task continues. After each round trip, the system will automatically switch the movement direction, cycling sequentially from the four directions: horizontal, vertical, 45° diagonal, to -45° diagonal. At the start of the task, the starting point of the stimulus ball will be randomly selected between -30° and 0° within the field of view, and this position will be used as the starting point. The stimulus ball will move towards the endpoint along the currently set direction (horizontal, vertical, or diagonal). The endpoint is offset by a maximum of 30° from the starting point, ensuring the ball remains within the field of vision, with its trajectory limited to -30° to 30°. The ball's speed remains constant, with an angular velocity of 16° / second. Users need to precisely track the ball's trajectory, maintaining a stable line of sight throughout its journey from the starting point to the endpoint. When the ball changes direction, users must quickly adjust their eye movements to maintain continuous focus on the target and complete a saccade task. This dynamic visual stimulation enhances the user's visual flexibility and accuracy, while strengthening the ability to quickly capture and track targets in different directions.

[0073] Specifically, such as Figure 4 As shown, in the visual tracking section of the visual training video, at the start of the task, four letter blocks (A, H, K, T) are displayed in the user's field of vision. They are arranged in a rectangle at the four corners of the 24° field of vision, and each letter block is approximately 1.72° in size (V-shaped optotype). In each round of the task, a letter block is randomly selected from A, H, K, and T, and it remains in the center for 2 seconds. The four surrounding letter blocks are randomly shuffled at the start of each round to increase the difficulty and avoid a fixed gaze pattern. The user's task is to quickly find the letter that matches the central letter block among the four surrounding letter blocks. By recognizing the central letter and matching it with the surrounding targets, the user can train their visual localization and dynamic reaction abilities.

[0074] In one specific embodiment, the location of visual field stimuli is determined using 24-2 visual field testing results. These results are used to assess visual field sensitivity within a central 24-degree range. A pattern deviation probability map is used to identify areas of visual acuity abnormality; if the pattern deviation probability map is unavailable due to poor visual field, a total probability deviation map is used as a substitute. Specifically, areas of visual acuity abnormality refer to points with a p-value less than 5% on the pattern deviation probability map (or total probability deviation map). Specifically, the small squares at these points contain markers reflecting a statistically significant decrease in visual acuity compared to the normal population.

[0075] The 24-2 visual field test results contained 54 fixed stimulus points distributed in a two-dimensional coordinate plane with the center of the visual field (0,0) as the origin, and the center interval between the points was 6°. The coordinates (angle_x, angle_y) of each point could be obtained from the report, and the coordinate lists (positions_r and positions_l) for the left and right eyes were defined respectively; each list contained 54 three-dimensional coordinates (angle_x, angle_y represent the visual coordinates, s indicates whether the position participated in training: 1 for participation, 0 for non-participation).

[0076] The `SpawnAndDestroyBall` function is located in the `VisualStimulation.cs` script. It is used to generate and destroy stimulus balls in the virtual scene. The position of the stimulus ball is determined by randomly selecting a location from a list of locations where `s=1`, ensuring that the stimulus generated by the ball is located within the visual field defect area. Each time a stimulus ball is generated, the system calculates its virtual coordinates based on the preset visual field angle and distance, and creates a stimulus ball instance in the scene. The stimulus ball is displayed in the scene for a predetermined time before being destroyed. After each destruction, the system automatically calculates the position of the next stimulus point and generates a new stimulus ball. This process is continuously repeated to simulate continuous stimulation in the visual field defect area. The pseudocode for the `SpawnAndDestroyBall` function is shown in Table 1. In Table 1, `ball03` is a template for dynamically generated spheres, defining the sphere's appearance, material, collider, and other properties. `ball03` is a pre-designed sphere (prefab) that is not visible in the scene, but multiple instances can be generated through code. In the code, `ball03` is used to dynamically generate new sphere instances using the `Instantiate()` method, similar to copying multiple spheres from this template. `ballparent` is a parent object (Container) used to organize and manage all dynamically generated spheres. Each generated sphere is set as a child object of `ballparent`, allowing for convenient and unified management of these generated spheres, and they all change with the position and rotation of `ballparent`. The `radius` is the same as the `radius` in the "Generate and Destroy Balls" section, which is the distance between the patient (camera plane) and the stimulus point (the plane containing the generated spheres).

[0077] Table 1

[0078]

[0079] During visual training, users wear a VR headset and watch personalized visual training videos on their smartphones. Training is conducted twice daily, with each session lasting 20 minutes, and each training cycle lasts 14 days. After 14 days, users need to undergo a visual acuity and visual field test at a hospital, including visual acuity, intraocular pressure, and 24-2 visual field testing, to assess the training effectiveness.

[0080] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0081] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A personalized visual training system for visual field defects based on virtual reality, characterized in that, The virtual reality-based personalized visual training system for visual field defects includes: head-mounted VR glasses, a mobile phone, a scene generation and video processing module, a head motion calculation module, a controller interaction module, and a cloud data analysis platform. The scene generation and video processing module is used to determine the visual field defect area based on the visual field detection results of the trainee; and to generate a visual training video based on the Unity3D platform according to the visual field defect area; the visual training video includes 3D training videos of different visual training tasks and left and right modes; the 3D training video includes: visual field stimulation instructions, right eye visual field stimulation section, rest section after removing the eye mask, left eye visual field stimulation section, rest section after removing the eye mask and closing the eyes, eye movement training instructions, saccade fixation section, visual tracking section, and visual matching section. The scene generation and video processing module is mounted on a cloud server and sends the generated visual training video to the visual training APP installed on the mobile phone. The VR headset is worn over the eyes of the person being trained; The head motion calculation module is used to acquire head acceleration data through the mobile phone accelerometer; and to perform gravity compensation and motion energy integration based on the head acceleration data. The handle interaction module is used to provide feedback via haptic feedback buttons; The cloud-based data analysis platform is connected to the head motion calculation module, the controller interaction module, the scene generation and video processing module, and the mobile phone; and is used to store head acceleration data and controller button event data, and generate a comprehensive evaluation report. During visual training, the trainee uses VR headsets to perform visual training based on a corresponding visual training video generated on their mobile phone. During this training, the head motion calculation module performs gravity compensation and motion energy integration based on head acceleration data. When the motion energy integration exceeds a threshold, the cloud-based data analysis platform pauses the visual training video on the mobile phone for a set time. Furthermore, based on the reaction time and success rate at different points determined by the controller button event data, the scene generation and video processing modules dynamically adjust the visual training video.

2. The personalized visual training system for visual field defects based on virtual reality according to claim 1, characterized in that, The sampling rate of the haptic feedback button is ≥100Hz.

3. The personalized visual training system for visual field defects based on virtual reality according to claim 1, characterized in that, The controller interaction module uses Bluetooth for interaction.

4. The personalized visual training system for visual field defects based on virtual reality according to claim 1, characterized in that, The scene generation and video processing module uses Adobe Premiere Pro for visual training video processing.

5. The personalized visual training system for visual field defects based on virtual reality according to claim 1, characterized in that, The scene generation and video processing module uses the SpawnAndDestroyBall function; the SpawnAndDestroyBall function is used to generate personalized stimuli for the visual stimulation part in the visual training video in a loop.

6. The personalized visual training system for visual field defects based on virtual reality according to claim 1, characterized in that, The scene generation and video processing module uses Visual Studio 2022 as the C# script compiler.

Images