Visual field promotion training system based on back flow visual channel motion space stimulation

Abstract

The invention discloses a visual field promotion training system based on back flow visual channel motion space stimulation, and relates to the technical field of visual field promotion. A visual field missing map management module is constructed and used for generating an initial visual field missing map of a user and baseline data of a visual field missing area; constructing a dynamic potential evaluation module, and generating a pixel-level visual potential map based on the baseline data; constructing a target generation module, and generating a dynamic pixel target for the visual potential map; constructing an eye movement tracking module which is used for tracking eyeball movement of the user in real time and confirming eye movement data of the user on the dynamic pixel target to obtain a training effect; a rehabilitation effect quantitative evaluation module is constructed, the training effect of completing the training task by the user is obtained, and a quantitative rehabilitation report is generated; through a closed-loop training mechanism combining dynamic potential evaluation and real-time eye movement tracking, a back flow visual pathway can be effectively activated, and personalized visual function rehabilitation training for a visual field defect area is realized.

Description

Technical Field

[0001] This invention relates to the field of visual field enhancement technology, specifically to a visual field enhancement training system based on backflow visual channel motion space stimulation. Background Technology

[0002] Visual field defects refer to the loss of visual perception in a certain area within the normal visual field. The causes of visual field defects are complex and diverse, mainly including glaucoma, retinal diseases, optic nerve diseases, etc. These diseases can directly damage the retina or optic nerve, resulting in a narrowing of the visual field or local visual field defects. Traditionally, visual impairment caused by binocular visual field defects is considered irreversible. However, recent studies have found that by activating visual residuals and utilizing the remaining cells, neurons, and the plasticity of binocular visual function, it may be possible to restore or reconstruct part of the visual field and visual function. However, existing visual field rehabilitation methods mainly rely on monocular visual stimulation training, such as visual recovery training. This method aims to activate residual visual function by presenting repetitive light stimulation in the boundary area of ​​the visual field defect. This type of method has significant drawbacks. First, its training mode is based on static or simple dynamic stimulation, failing to effectively mobilize the unique function of the backflow visual channel in processing spatial information of motion, thus limiting the depth of neural plasticity induction. The current technology lacks precise personalized assessment and control mechanisms, often employing uniform stimulation parameters and training difficulty, resulting in insufficient training targeting, low patient participation, and difficulty in forming an effective adaptive rehabilitation loop. Furthermore, traditional methods rely heavily on standard perimeter examinations for visual field assessment, failing to identify and utilize residual visual areas with recovery potential at the pixel level, leading to unclear training targets and limited efficacy. Although neuroscience research has confirmed the plasticity of the brain's visual system and that activation of residual visual areas is key to visual field function recovery, current rehabilitation technologies have not yet been able to translate these theoretical breakthroughs into efficient and precise clinical training programs. Therefore, it is of great significance to develop a visual field enhancement training system based on backflow visual channel motion space stimulation. Summary of the Invention

[0003] The purpose of this invention is to provide a visual field enhancement training system based on backflow visual channel motion space stimulation to address the shortcomings of the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a visual field enhancement training system based on backflow visual channel motion space stimulation, comprising: The field-of-view missing map management module is used to import and digitize the user's initial field-of-view missing map and establish baseline data for the field-of-view missing area. Dynamic potential assessment module: Connected to the view gap map management module, it performs visual potential assessment tasks based on baseline data and generates pixel-level visual potential maps; Target generation module: Connected to the dynamic potential assessment module, it generates dynamic pixel targets for the visual potential map; Eye-tracking module: Connected to the target generation module, it is used to track the user's eye movements in real time and confirm that the user's eye movement data on the dynamic pixel target has achieved the training effect; Quantitative assessment module for rehabilitation effect: Connects with the eye-tracking module to obtain the training effect of the user's completion of training tasks and compares it with baseline data to generate a quantitative rehabilitation report.

[0005] In a preferred embodiment, the view gap map management module includes: The data interface unit is used to receive standard format field of view data files from external field of view inspection equipment; The map parsing unit is used to convert the view data file into a digital view map and identify the pixel coordinates of the view defect area as baseline data.

[0006] In a preferred embodiment, the dynamic potential assessment module includes: The stimulus presentation unit is used to generate a first optical flow particle as a guide point in the user's non-missing visual field area, and control the first optical flow particle to move to a candidate position in the visual field defect area, wherein the candidate position is a random pixel in the visual field defect area, and each pixel is selected only once. The saccade analysis unit is used to capture the user's eye movement data during the process of the first optical flow particle moving to the candidate position through the eye tracking module. If the eye movement data shows that the user has an autonomous saccade eye movement pointing to the candidate position, the candidate position is determined to have visual potential, and it is included as a potential pixel in the visual potential map.

[0007] In a preferred embodiment, the target generation module includes: The target generation unit is used to set dynamic optical flow particles as dynamic pixel targets at non-field defect locations. The mode management unit is used to configure different binocular co-display modes for dynamic pixel targets.

[0008] In a preferred embodiment, the mode management unit includes: The dual-eye collaborative display modes include contrast balance mode, color complementary mode, and stereo depth mode; In contrast balance mode, the dynamic pixel target seen by the left and right eyes is in the same position and direction of movement, but the display contrast is different. The color complementary mode breaks down the dynamic pixel target into different color components and presents them to the left and right eyes respectively. The stereo depth mode generates a virtual sense of depth for the dynamic pixel target by adjusting the horizontal parallax of the pixels in the dynamic pixel target.

[0009] In a preferred embodiment, the eye-tracking module includes: The training image generation unit imports the digital field of view map into the 3D split-view device to generate a training base map, and generates a training image by generating a dynamic pixel target with a preset mode in the training base map, wherein the preset mode includes at least one binocular collaborative display mode. The target path planning unit is used to control the dynamic pixel target to move globally along a preset path into the area of ​​missing vision. The stimulus enhancement unit executes a stimulus enhancement mechanism when a pixel in the dynamic pixel target coincides with a pixel in the visual potential map. The stimulus enhancement mechanism enhances the optical parameters of the pixel, including brightness, contrast, color saturation, and flicker frequency. The eye-tracking unit acquires the user's eye movement data in real time when the dynamic pixel target enters the visual field defect area. The eye-tracking data includes gaze point coordinates, timestamps, and behavioral event data. The timestamps include the start and end times of the gaze, and the behavioral event data includes saccade events, fixation events, and smooth tracking events. The data transmission unit is used to transmit eye-tracking data to the 3D split-view device in real time. In the training unit, users view training images through a 3D split-view device. Users focus their gaze on a dynamic pixel target according to voice prompts, and their gaze follows the movement of the dynamic pixel target. For a specific pixel of a dynamic pixel target in a visual field defect area, if the user's eye movement data meets the preset eye movement data, the eye movement data of that pixel is transmitted to the 3D split-view device through the data transmission unit. After receiving the eye movement data, the 3D split-view device performs an extinguishing operation on the pixel corresponding to the eye movement data. The number of pixels that went out during each training session is recorded as the training result.

[0010] In a preferred embodiment, the data transmission unit includes: The channel mapping table management unit is used to establish and maintain a permission mapping map corresponding to the pixel coordinates of the training image, and store the permission mapping map in the 3D split-view device; The dynamic channel allocation unit allocates a dynamic transmission channel to each pixel based on the permission mapping map; The pixel coordinates of the dynamic pixel target that enters the area of ​​missing vision are located, and an advanced index is established at the corresponding coordinates in the permission mapping map. Based on the advanced index, a high-priority dynamic transmission channel is built between the 3D split-view device and the corresponding pixel. The motion path of the dynamic pixel target in the field of view defect area is obtained, the coordinates of the pixel points that will pass through within a preset time in the future are obtained, and intermediate indexes are established with the corresponding coordinates in the permission mapping map. Based on the intermediate indexes, a medium-priority dynamic transmission channel is constructed between the 3D split view device and the corresponding pixel points. The remaining pixels establish a low-priority dynamic transmission channel with the 3D split-view device.

[0011] In a preferred embodiment, the rehabilitation effect quantitative assessment module includes: The potential change analysis unit is used to compare the visual potential map generated by the re-evaluation after training with the baseline data before training, and to calculate the rate of change of the number of potential pixels and the change of spatial distribution in the visual field missing area. The training performance statistics unit is used to count the number of pixels successfully turned off by the user during the training process, and to count the extinguishing rate of pixels according to different binocular co-display modes. The extinguishing rate is the ratio of the number of extinguished pixels to the number of pixels in the dynamic pixel target. The integrated report generation unit integrates the outputs of the potential change analysis unit and the training effectiveness statistics unit to generate a quantitative rehabilitation report that includes a visual field recovery trend map, pattern fitness analysis, and training suggestions.

[0012] The technical effects and advantages provided by the present invention in the above technical solution are as follows: 1. This invention effectively improves the neural efficiency of rehabilitation training for patients with visual field defects by constructing an active training closed loop of assessment-interaction-feedback. The system accurately quantifies the residual function of the visual field defect area through visually guided saccade tasks and generates a visual potential map as a training benchmark. During the training process, patients need to interact with dynamic targets through precise eye movement tracking and gaze persistence mechanisms to successfully activate the spatial positioning and motion analysis functions of the backflow visual channel. Based on this task-driven active training mode, it is more in line with the neural plasticity mechanism and can effectively strengthen the repair and reconstruction of damaged visual pathways. At the same time, real-time interactive feedback significantly improves the patient's participation and training continuity. 2. This invention constructs a highly intelligent training system by integrating content-aware data transmission and a closed-loop adaptive control scheme. The system establishes a dynamic mapping between visual space and transmission channels, providing dedicated transmission channels for key interactive data and ensuring that real-time feedback latency is below the perception threshold. Simultaneously, based on the deep integration of real-time training performance and visual potential maps, the system can dynamically adjust target parameters and training difficulty to achieve efficient personalized training programs. Based on this closed-loop adaptive mechanism, not only is the effectiveness of a single training session guaranteed, but continuous quantitative evaluation and dynamic optimization also ensure that rehabilitation effects accumulate and improve with the training process, forming a scientific rehabilitation progression curve. Attached Figure Description

[0013] 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 recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0014] Figure 1 This is a system flowchart of the present invention.

[0015] Figure 2 This is a logic block diagram of the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1, please refer to Figure 1 and Figure 2 As shown in this embodiment, a visual field enhancement training system based on backflow visual channel motion space stimulation includes: The field-of-view missing map management module is used to import and digitize the user's initial field-of-view missing map and establish baseline data for the field-of-view missing area. Dynamic potential assessment module: Connected to the view gap map management module, it performs visual potential assessment tasks based on baseline data and generates pixel-level visual potential maps; Target generation module: Connected to the dynamic potential assessment module, it generates dynamic pixel targets for the visual potential map; Eye-tracking module: Connected to the target generation module, it is used to track the user's eye movements in real time and confirm that the user's eye movement data on the dynamic pixel target has achieved the training effect; Quantitative assessment module for rehabilitation effect: Connects with the eye-tracking module to obtain the training effect of the user's completion of training tasks and compares it with baseline data to generate a quantitative rehabilitation report; Furthermore, visual field defects refer to the phenomenon of visual perception loss in a certain area within the normal visual field. The causes of visual field defects are complex and diverse, mainly including glaucoma, retinal diseases, optic nerve diseases, etc. These diseases can directly damage the retina or optic nerve, resulting in a narrowing of the visual field or local visual field loss. Traditionally, visual impairment caused by binocular visual field defects is considered irreversible. However, recent studies have found that by activating visual residuals and utilizing the remaining cells, neurons, and the plasticity of binocular visual function, it may be possible to restore or reconstruct part of the visual field and visual function. However, existing visual field rehabilitation methods mainly rely on monocular visual stimulation training, such as visual recovery training. This method aims to activate residual visual function by presenting repetitive light stimulation in the boundary area of ​​the visual field defect. This type of method has significant drawbacks. First, its training mode is based on static or simple dynamic stimulation, failing to effectively mobilize the unique function of the backflow visual channel in processing spatial information of motion, thus limiting the depth of neural plasticity induction. The current technology lacks precise personalized assessment and control mechanisms, often employing uniform stimulation parameters and training difficulty, resulting in insufficient training targeting, low patient participation, and difficulty in forming an effective adaptive rehabilitation loop. Furthermore, traditional methods rely heavily on standard perimeter examinations for visual field assessment, failing to identify and utilize residual visual areas with recovery potential at the pixel level, leading to unclear training targets and limited efficacy. Although neuroscience research has confirmed the plasticity of the brain's visual system and that activation of residual visual areas is key to visual field function recovery, current rehabilitation technologies have not yet been able to translate these theoretical breakthroughs into efficient and precise clinical training programs. This invention effectively improves the neural efficiency of rehabilitation training for patients with visual field defects by constructing an active training closed loop of assessment-interaction-feedback. The system accurately quantifies the residual function of the visual field defect area through visually guided saccade tasks and generates a visual potential map as a training benchmark. During the training process, patients need to interact with dynamic targets through precise eye movement tracking and gaze persistence mechanisms to successfully activate the spatial positioning and motion analysis functions of the backflow visual channel. This task-driven active training mode is more in line with the neural plasticity mechanism and can effectively strengthen the repair and reconstruction of damaged visual pathways. At the same time, real-time interactive feedback significantly improves patient participation and training continuity. By integrating content-aware data transmission and closed-loop adaptive control, a highly intelligent training system is constructed. The system establishes a dynamic mapping between visual space and transmission channels, providing dedicated transmission channels for key interactive data and ensuring that real-time feedback latency is below the perception threshold. Simultaneously, based on the deep integration of real-time training performance and visual potential maps, the system can dynamically adjust target parameters and training difficulty to achieve efficient personalized training programs. This closed-loop adaptive mechanism not only ensures the effectiveness of individual training sessions but also ensures that rehabilitation effects accumulate and improve with the training process through continuous quantitative evaluation and dynamic optimization, forming a scientific rehabilitation progression curve.

[0018] In one embodiment, the view gap map management module includes: The data interface unit is used to receive standard format field of view data files from external field of view inspection equipment; The map parsing unit is used to convert the field data file into a digital field map and identify the pixel coordinates of the field defect area as baseline data. Furthermore, the visual field defect map management module establishes a communication connection with external visual field examination equipment through a data interface unit. This unit supports parsing visual field data files that conform to medical digital imaging and communication standards, and can automatically extract the visual field sensitivity matrix, test point coordinate mapping relationship, and individual patient calibration parameters contained in the file. The data interface unit adopts an asynchronous transmission mechanism to ensure the complete reception of large-capacity visual field data and verifies the data integrity through a verification algorithm. The map parsing unit then converts the received structured visual field data into a digital visual field map with a preset resolution. This process includes coordinate system unification processing, grayscale interpolation calculation, and spatial grid mapping. Based on a preset visual field defect judgment threshold, the system automatically identifies test point areas with sensitivity below the critical value, accurately delineates the contour boundary of the visual field defect area through an edge detection algorithm, and records the coordinates of all pixels in these areas and their sensitivity values ​​as baseline data. Based on this process, a fully automatic conversion from the original visual field report to the digital defect map is realized, providing an accurate spatial benchmark for subsequent visual potential assessment and training.

[0019] In one embodiment, the dynamic potential assessment module includes: The stimulus presentation unit is used to generate a first optical flow particle as a guide point in the user's non-missing visual field area, and control the first optical flow particle to move to a candidate position in the visual field defect area, wherein the candidate position is a random pixel in the visual field defect area, and each pixel is selected only once. The saccade analysis unit is used to capture the eye movement data of the user during the process of the first optical flow particle moving to the candidate position through the eye tracking module. If the eye movement data shows that the user has an autonomous saccade eye movement pointing to the candidate position, the candidate position is determined to have visual potential, and it is included as a potential pixel in the visual potential map. Furthermore, the stimulus presentation unit of the dynamic potential assessment module first generates a first optical flow particle as a guide marker within the user's intact visual field. This particle enhances its salience through high-frequency micro-twitch, for example, by flashing brightness at a frequency of 15 to 20 Hz. Subsequently, the system controls the particle to move along a smooth Bézier curve trajectory to a preset candidate position within the visual field defect area. The candidate positions are uniformly selected within the defect area using a hierarchical random sampling algorithm. This algorithm divides the visual field defect area into multiple grids based on a preset resolution to ensure uniform spatial coverage. The system also maintains a list of visited coordinates to ensure that each pixel is selected only once throughout the entire assessment process. During the movement of the first optical flow particle, the saccade analysis unit continuously collects the user's eye movement data through the eye-tracking module. The system processes the collected raw eye movement signals in real time. First, a velocity-based event detection algorithm is used to identify the start and end of saccade movements. When a saccade event is detected, the system extracts the motion of that event. The system learns features, such as comparing the peak velocity of the saccade with a velocity threshold set based on the user's baseline eye-tracking data or population norm data, to distinguish between voluntary saccades and unconscious micro-nystagmus. Simultaneously, the system calculates the motion vector direction of the saccade event and geometrically compares it with the ideal vector direction of the first optical flow particle moving to the candidate position. The angle between the two must be less than the angle tolerance range set according to eye-tracking accuracy and training requirements. Finally, it determines whether the endpoint of the saccade falls within a spatial tolerance window centered on the candidate position. In the process of determining potential pixels, the system only determines that the user has perceived the stimulus at the candidate position and generated an effective voluntary saccade behavior when the three feature parameters of the saccade event—peak velocity, motion direction, and endpoint position—simultaneously meet their respective judgment criteria. This indicates that the candidate position has visual perception potential. The pixel coordinates are then binary-coded and marked in a visual potential map, which clearly displays the positions of all pixels with visual potential in a spatially distributed dot matrix format.

[0020] In one embodiment, the target generation module includes: The target generation unit is used to set dynamic optical flow particles as dynamic pixel targets at non-field defect locations. A mode management unit is used to configure different binocular collaborative display modes for dynamic pixel targets; Furthermore, the target generation unit initializes a set of dynamic optical flow particles as dynamic pixel targets in the non-defective area of ​​the user's field of vision. These particles generate coherent motion patterns through a preset algorithm, forming a visual stimulus flow with a clear direction of motion. The algorithm first initializes a particle emission source in the non-visual defect area. Each particle represents an independent optical stimulus unit, i.e., a pixel. The algorithm controls the trajectory of the particles through a vector field. This vector field can be configured as a basic motion pattern such as radial diffusion, vortex rotation, or directional translation. The particle motion follows the principle of coherence in physical simulation, and the motion vector between adjacent particles maintains a smooth transition, thereby forming an optical flow field with overall consistency. At the same time, the particle system also includes a complete life cycle management mechanism. After generation, each particle undergoes brightness gradation and size change according to a preset life cycle, and is automatically regenerated at the emission source at the end of its life cycle, thereby maintaining the continuity and dynamic balance of visual stimulation. The mode management unit configures different binocular co-display modes for these dynamic pixel targets according to training requirements.

[0021] In one embodiment, the mode management unit includes: The dual-eye collaborative display modes include contrast balance mode, color complementary mode, and stereo depth mode; In contrast balance mode, the dynamic pixel target seen by the left and right eyes is in the same position and direction of movement, but the display contrast is different. The color complementary mode breaks down the dynamic pixel target into different color components and presents them to the left and right eyes respectively. The stereo depth mode generates a virtual sense of depth for the dynamic pixel target by adjusting the horizontal parallax of the pixels in the dynamic pixel target. Furthermore, the mode management unit implements three binocular collaborative modes through split-view display technology. In contrast balance mode, the system generates target images with the same spatial coordinates and motion vectors for both eyes, but independently adjusts their grayscale dynamic range. Specifically, this is achieved by adjusting the gain and offset parameters of their respective image signals in real time, so that the dominant eye receives low-contrast stimulation while the inferior eye receives high-contrast stimulation. In color complementary mode, the system decomposes the original color of the target into two complementary color components. For example, a white target is decomposed into a cyan component displayed to the left eye and a red component displayed to the right eye. These components are then fused in the visual center through the principle of color superposition to form a white light perception. In stereo depth mode, the system uses a parallax rendering engine to calculate the horizontal displacement of each pixel in the left and right eye views in real time. Based on the target virtual depth value, it generates binocular image pairs with subtle horizontal position differences. When these images are projected onto the corresponding eyeballs through a stereo display device, the visual cortex of the brain automatically fuses this parallax information and generates a three-dimensional spatial perception with a clear front-back hierarchical relationship. All mode parameters can be dynamically optimized and adjusted according to the real-time monitoring of binocular competition. Among them, the contrast balance mode is designed based on binocular inhibition and competition mechanisms. In neurological diseases such as glaucoma, the weaker eye often has lower signal quality. The dominant eye suppresses the weaker eye, leading to deeper inhibition and inability to activate visual potential. This mode dynamically reduces the contrast of the dominant eye (e.g., to 30%) while simultaneously increasing the contrast of the weaker eye (e.g., to 100%), forcing both eyes to work together. This contrast difference can significantly improve the patient's inhibition depth and balanced contrast, helping to rebuild binocular balance and activate residual visual pathways. The color complement mode utilizes the separation characteristics of the temporal and dorsal pathways in color vision. By stimulating different colors separately, different visual channels are activated. For example, decomposing the target into short-wavelength cyan in the left eye and long-wavelength red in the right eye can induce stronger visual stimulation in the visual cortex. Color antagonism and fusion response enhance the saliency of visual signals and the ability to capture attention. This color-separated presentation not only improves the detection threshold of the target in the visual field defect area, but also enhances the activation of neural circuits related to object recognition in the ventral pathway through color contrast enhancement, thereby helping to improve the patient's ability to identify and track dynamic targets. The stereo depth mode is based on the dorsal pathway mechanism of binocular disparity and depth perception. By fine-tuning the horizontal displacement of the target in the left and right eye diagrams, a virtual three-dimensional motion trajectory is constructed. For example, by setting the target to oscillate back and forth along the Z-axis, combined with changes in horizontal disparity, the MT / V5 region in the dorsal pathway responsible for spatial localization analysis can be activated.This mode is particularly suitable for glaucoma patients with impaired stereoscopic vision. It promotes binocular fusion and spatial perception through dynamic depth stimulation. In actual training, the system can adaptively adjust the disparity amplitude and motion speed based on the patient's real-time stereoscopic detection results, such as the wide-range stereoscopic vision level and fine stereoscopic vision registration, to achieve a personalized stimulation plan. Furthermore, the switching of this mode can be implemented independently. Patients can project the training module onto a regular display screen and train using a simple split vision device, so they can conduct stimulation training at home without the need for a hospital training environment.

[0022] In one embodiment, the eye-tracking module includes: The training image generation unit imports the digital field of view map into the 3D split-view device to generate a training base map, and generates a training image by generating a dynamic pixel target with a preset mode in the training base map, wherein the preset mode includes at least one binocular collaborative display mode. The target path planning unit is used to control the dynamic pixel target to move globally along a preset path into the area of ​​missing vision. The stimulus enhancement unit executes a stimulus enhancement mechanism when a pixel in the dynamic pixel target coincides with a pixel in the visual potential map. The stimulus enhancement mechanism enhances the optical parameters of the pixel, including brightness, contrast, color saturation, and flicker frequency. The eye-tracking unit acquires the user's eye movement data in real time when the dynamic pixel target enters the visual field defect area. The eye-tracking data includes gaze point coordinates, timestamps, and behavioral event data. The timestamps include the start and end times of the gaze, and the behavioral event data includes saccade events, fixation events, and smooth tracking events. The data transmission unit is used to transmit eye-tracking data to the 3D split-view device in real time. In the training unit, users view training images through a 3D split-view device. Users focus their gaze on a dynamic pixel target according to voice prompts, and their gaze follows the movement of the dynamic pixel target. For a specific pixel of a dynamic pixel target in a visual field defect area, if the user's eye movement data meets the preset eye movement data, the eye movement data of that pixel is transmitted to the 3D split-view device through the data transmission unit. After receiving the eye movement data, the 3D split-view device performs an extinguishing operation on the pixel corresponding to the eye movement data. The number of pixels that went out during each training session was recorded as the training effect. Furthermore, the training image generation unit of the eye-tracking module first aligns the digitized visual field map with the visual potential map and then sends it to the graphics rendering engine to generate a training base map containing background visual field information. Then, a dynamic pixel target with a specific binocular co-display mode is superimposed on the training base map according to the output configuration of the mode management unit to generate a training image. The target path planning unit uses a trajectory generation algorithm based on B-spline curves to plan a smooth motion path starting from the intact visual field area and meandering through the entire visual field defect area, ensuring that the target can systematically pass through all identified visual potential points. The stimulus reinforcement unit compares the coordinate mapping relationship between the target's current position and the visual potential map in real time. When a target is detected... When a pixel coincides with a potential pixel, a multi-parameter enhancement protocol is immediately activated. This simultaneously increases the pixel's luminous intensity, adjusts its contrast with the background, enhances color purity, and superimposes a pulsed flashing effect at a specific frequency. The steps for determining if a dynamic pixel target coincides with a potential pixel are as follows: The system internally generates an index table pointing to the current target path point. At each display frame refresh, the algorithm calculates the direction vector of the dynamic pixel target moving to the next path point at a constant speed and updates the current pixel coordinates accordingly. When the dynamic pixel target's pixel coordinates coincide with the target path point coordinates, the system automatically increments the index to the next path point, thereby driving the target to move gradually along a predetermined trajectory. The stimulation enhancement unit then updates the index at this location. Immediately after the scan, these real-time coordinates are compared with pre-stored coordinates in the visual potential map to determine overlap. The gaze tracking unit uses an eye tracker based on the corneal reflection principle, continuously capturing eye image sequences at a sampling frequency of over 15,000 times per minute using a built-in infrared camera. The system performs pupil contour recognition and corneal reflection spot localization in real time for each frame. By calculating the two-dimensional vector formed by the relative positions of the two, it calculates the gaze point coordinates accurate to the pixel level in the screen coordinate system. These coordinates represent the instantaneous focal position of the user's gaze on the display plane. In terms of behavioral event data recording, the system continuously analyzes the gaze point coordinate sequence with millisecond-level timestamps through a multi-level event detection engine. First, it calculates the eye movement between adjacent sampling points. The system uses motion vectors and instantaneous velocities. When a velocity exceeds a threshold determined by statistical analysis of a large amount of eye-tracking data, it is marked as the start point of a saccade event and tracked continuously until the velocity falls back to a stopping threshold. The start and end timestamps, motion trajectory, and peak velocity parameters of the event are recorded as saccade events. For stable coordinate point clusters with velocities consistently below a specific threshold, they are identified as fixation events, and their average gaze coordinates and duration are recorded. When the eye-tracking velocity vector is detected to be synchronized with the target's motion trajectory and the directional deviation is less than a set angle, it is identified as a smooth following event, and its motion matching parameters are recorded. All parsed event data are indexed by timestamps and together with the corresponding gaze point coordinates, they form a complete structured eye-tracking data stream.Users view training images through a 3D split-view device and attempt to focus their gaze and follow the movement of a dynamic pixel target based on voice prompts. The system executes the following judgment logic: when the dynamic pixel target is located within a visual field defect area and the eye movement data acquired by the gaze tracking unit meets preset conditions, the eye movement data is transmitted to the 3D split-view device in real time via the data transmission unit. The 3D split-view device sends an extinguishing request to the pixel corresponding to the eye movement data. The pixel accepts the extinguishing request and performs an extinguishing operation, instantly reducing the optical output intensity of the pixel to a minimum, making it visually disappear or nearly invisible. The preset conditions are: for saccade events, verifying that its peak velocity exceeds a threshold and the landing point points to the target; for fixation events, confirming that its gaze point remains within the tolerance range around the target for more than the minimum time; for smooth following events, evaluating that its motion trajectory matches the target path better than a set standard. The system sets initial thresholds for various events based on general eye movement norm data, and the velocity threshold for saccade events is... The initial thresholds are set within a typical range based on eye-tracking statistics from a large number of healthy individuals. The minimum duration threshold for fixation events is initialized based on the minimum time required to ensure visual cognitive processing. The path matching tolerance threshold for smooth tracking is configured based on the target's movement speed and the expected tracking accuracy. Furthermore, before formal training begins, the system performs a brief calibration process by having the user track a set of targets with known paths to measure their individualized baseline eye-tracking characteristics, such as average saccadic velocity and fixation stability. The system then fine-tunes the initial thresholds based on this personalized baseline data. During training, the system dynamically relaxes or tightens certain thresholds based on the user's real-time performance. For example, after a user successfully tracks multiple times consecutively, the system appropriately increases the requirement for smooth tracking path matching to achieve more effective stimulus training. Finally, the system records the total number of pixels successfully extinguished throughout the entire training process as a core indicator for quantitatively evaluating the training effect.

[0023] In one embodiment, the data transmission unit includes: The channel mapping table management unit is used to establish and maintain a permission mapping map corresponding to the pixel coordinates of the training image, and store the permission mapping map in the 3D split-view device; The dynamic channel allocation unit allocates a dynamic transmission channel to each pixel based on the permission mapping map; The pixel coordinates of the dynamic pixel target that enters the area of ​​missing vision are located, and an advanced index is established at the corresponding coordinates in the permission mapping map. Based on the advanced index, a high-priority dynamic transmission channel is built between the 3D split-view device and the corresponding pixel. The motion path of the dynamic pixel target in the field of view defect area is obtained, the coordinates of the pixel points that will pass through within a preset time in the future are obtained, and intermediate indexes are established with the corresponding coordinates in the permission mapping map. Based on the intermediate indexes, a medium-priority dynamic transmission channel is constructed between the 3D split view device and the corresponding pixel points. The remaining pixels establish a low-priority dynamic transmission channel with the 3D split-view device; Furthermore, the channel mapping table management unit of the data transmission unit creates and maintains a permission mapping map that completely corresponds to the resolution of the training image. Essentially, it is a permission mapping matrix, where the value of each element represents the data transmission priority level of its corresponding screen pixel. The dynamic channel allocation unit then schedules network resources in real time based on this matrix. The steps are as follows: the system continuously receives real-time coordinate streams from the target path planning unit. When any pixel of the dynamic pixel target is detected entering a missing area of ​​the field of view, these coordinate positions are immediately locked in the permission mapping matrix and marked as the highest priority. Simultaneously, a dedicated real-time transmission channel is allocated to these coordinates. This channel uses a pre-defined... The system employs bandwidth-saving and minimum-latency routing protocols to ensure the immediate delivery of eye-tracking data. It also uses a trajectory tracking algorithm to monitor the target's movement in real time and calculates the sequence of pixel coordinates that will be passed in the very short future based on the known global motion path. These coordinates are then pre-marked as medium priority in the permission mapping matrix and configured with transmission channels that offer high-quality service but allow for microsecond-level latency. All other pixels not marked as high or medium priority in the mapping matrix are assigned lower-priority dynamic transmission channels. Simultaneously, all dynamic transmission channels use timestamped data packet sequences for transmission, ensuring strict synchronization between eye-tracking data and visual stimuli.

[0024] In one embodiment, the quantitative assessment module for rehabilitation effectiveness includes: The potential change analysis unit is used to compare the visual potential map generated by the re-evaluation after training with the baseline data before training, and to calculate the rate of change of the number of potential pixels and the change of spatial distribution in the visual field missing area. The training performance statistics unit is used to count the number of pixels successfully turned off by the user during the training process, and to count the extinguishing rate of pixels according to different binocular co-display modes. The extinguishing rate is the ratio of the number of extinguished pixels to the number of pixels in the dynamic pixel target. The integrated report generation unit is used to integrate the output results of the potential change analysis unit and the training efficacy statistics unit to generate a quantitative rehabilitation report that includes a visual field recovery trend map, pattern fitness analysis and training suggestions. Furthermore, by performing a precise pixel-level spatial registration, the newly generated visual potential map after training is overlaid and compared with the original digitized visual field map and the initial visual potential map stored before training. This unit first calculates the rate of change in the total number of potential pixels marked as potential points in the visual field missing area, that is, the percentage difference between the number of potential points in the visual potential map after training and the number of potential points in the initial visual potential map before training. At the same time, an algorithm based on grid density analysis is used to quantify the change in spatial distribution. By dividing the entire visual field into uniform micro-grids and counting the number of pixels marked as potential points in each grid, the correlation coefficient of grid density distribution before and after training and the centroid offset distance of high-density areas are calculated. The training performance statistics unit extracts the original records of each training session from the system log and analyzes the data according to three binocular co-display modes: contrast balance mode, color complementary mode, and stereo depth mode. The system performs statistical analysis, calculating the training efficiency index for each mode: the ratio of the total number of successfully extinguished pixels in that mode to the total number of pixels of the dynamic pixel target passing through the visual field loss area in that mode. This yields the specific extinguishing rate for each mode. The comprehensive report generation unit receives these analysis results and uses a data visualization engine to plot the change rate of potential point count over time as a visual field recovery trend chart. It also displays the extinguishing rates of different modes as radar charts to form a mode fitness analysis. Based on these data fusion results, the system executes decision-making logic to automatically generate personalized training suggestions. The core rules of this logic include: comparing the extinguishing rate values ​​of different binocular co-display modes and automatically recommending the mode with the highest extinguishing rate as the priority training mode for the next stage; identifying and locating visual field areas where the change rate of potential point count is below a set threshold, and suggesting increasing the intensity of training stimulation or adjusting the target's movement path in that area to improve stimulation coverage. Finally, the system integrates all quantitative charts and rule-based training suggestions to form a structured quantitative rehabilitation report.

[0025] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A visual field enhancement training system based on backflow visual channel motion space stimulation, characterized in that, The field-of-view missing map management module is used to import and digitize the user's initial field-of-view missing map and establish baseline data for the field-of-view missing area. Dynamic potential assessment module: Connected to the view gap map management module, it performs visual potential assessment tasks based on baseline data and generates pixel-level visual potential maps; Target generation module: Connected to the dynamic potential assessment module, it generates dynamic pixel targets for the visual potential map; Eye-tracking module: Connected to the target generation module, it is used to track the user's eye movements in real time and confirm that the user's eye movement data on the dynamic pixel target has achieved the training effect; Quantitative assessment module for rehabilitation effect: Connects with the eye-tracking module to obtain the training effect of the user's completion of training tasks and compares it with baseline data to generate a quantitative rehabilitation report.

2. The visual field enhancement training system based on backflow visual channel motion space stimulation according to claim 1, characterized in that, The missing field map management module includes: The data interface unit is used to receive standard format field of view data files from external field of view inspection equipment; The map parsing unit is used to convert the view data file into a digital view map and identify the pixel coordinates of the view defect area as baseline data.

3. The visual field enhancement training system based on backflow visual channel motion space stimulation according to claim 1, characterized in that, The dynamic potential assessment module includes: The stimulus presentation unit is used to generate a first optical flow particle as a guide point in the user's non-missing visual field area, and control the first optical flow particle to move to a candidate position in the visual field defect area, wherein the candidate position is a random pixel in the visual field defect area, and each pixel is selected only once. The saccade analysis unit is used to capture the user's eye movement data during the process of the first optical flow particle moving to the candidate position through the eye tracking module. If the eye movement data shows that the user has an autonomous saccade eye movement pointing to the candidate position, the candidate position is determined to have visual potential, and it is included as a potential pixel in the visual potential map.

4. The visual field enhancement training system based on backflow visual channel motion space stimulation according to claim 1, characterized in that, The target generation module includes: The target generation unit is used to set dynamic optical flow particles as dynamic pixel targets at non-field defect locations. The mode management unit is used to configure different binocular co-display modes for dynamic pixel targets.

5. The visual field enhancement training system based on backflow visual channel motion space stimulation according to claim 4, characterized in that, The mode management unit includes: The dual-eye collaborative display modes include contrast balance mode, color complementary mode, and stereo depth mode; In contrast balance mode, the dynamic pixel target seen by the left and right eyes is in the same position and direction of movement, but the display contrast is different. The color complementary mode breaks down the dynamic pixel target into different color components and presents them to the left and right eyes respectively. The stereo depth mode generates a virtual sense of depth for the dynamic pixel target by adjusting the horizontal parallax of the pixels in the dynamic pixel target.

6. The visual field enhancement training system based on backflow visual channel motion space stimulation according to claim 5, characterized in that, The eye-tracking module includes: The training image generation unit imports the digital field of view map into the 3D split-view device to generate a training base map, and generates a training image by generating a dynamic pixel target with a preset mode in the training base map, wherein the preset mode includes at least one binocular collaborative display mode. The target path planning unit is used to control the dynamic pixel target to move globally along a preset path into the area of ​​missing vision. The stimulus enhancement unit executes a stimulus enhancement mechanism when a pixel in the dynamic pixel target coincides with a pixel in the visual potential map. The stimulus enhancement mechanism enhances the optical parameters of the pixel, including brightness, contrast, color saturation, and flicker frequency. The eye-tracking unit acquires the user's eye movement data in real time when the dynamic pixel target enters the visual field defect area. The eye-tracking data includes gaze point coordinates, timestamps, and behavioral event data. The timestamps include the start and end times of the gaze, and the behavioral event data includes saccade events, fixation events, and smooth tracking events. The data transmission unit is used to transmit eye-tracking data to the 3D split-view device in real time. In the training unit, users view training images through a 3D split-view device. Users focus their gaze on a dynamic pixel target according to voice prompts, and their gaze follows the movement of the dynamic pixel target. For a specific pixel of a dynamic pixel target in a visual field defect area, if the user's eye movement data meets the preset eye movement data, the eye movement data of that pixel is transmitted to the 3D split-view device through the data transmission unit. After receiving the eye movement data, the 3D split-view device performs an extinguishing operation on the pixel corresponding to the eye movement data. The number of pixels that went out during each training session is recorded as the training result.

7. The visual field enhancement training system based on backflow visual channel motion space stimulation according to claim 6, characterized in that, The data transmission unit includes: The channel mapping table management unit is used to establish and maintain a permission mapping map corresponding to the pixel coordinates of the training image, and store the permission mapping map in the 3D split-view device; The dynamic channel allocation unit allocates a dynamic transmission channel to each pixel based on the permission mapping map; The pixel coordinates of the dynamic pixel target that enters the area of ​​missing vision are located, and an advanced index is established at the corresponding coordinates in the permission mapping map. Based on the advanced index, a high-priority dynamic transmission channel is built between the 3D split-view device and the corresponding pixel. The motion path of the dynamic pixel target in the field of view defect area is obtained, the coordinates of the pixel points that will pass through within a preset time in the future are obtained, and intermediate indexes are established with the corresponding coordinates in the permission mapping map. Based on the intermediate indexes, a medium-priority dynamic transmission channel is constructed between the 3D split view device and the corresponding pixel points. The remaining pixels establish a low-priority dynamic transmission channel with the 3D split-view device.

8. The visual field enhancement training system based on backflow visual channel motion space stimulation according to claim 1, characterized in that, The quantitative evaluation module for rehabilitation effectiveness includes: The potential change analysis unit is used to compare the visual potential map generated by the re-evaluation after training with the baseline data before training, and to calculate the rate of change of the number of potential pixels and the change of spatial distribution in the visual field missing area. The training performance statistics unit is used to count the number of pixels successfully turned off by the user during the training process, and to count the extinguishing rate of pixels according to different binocular co-display modes. The extinguishing rate is the ratio of the number of extinguished pixels to the number of pixels in the dynamic pixel target. The integrated report generation unit integrates the outputs of the potential change analysis unit and the training effectiveness statistics unit to generate a quantitative rehabilitation report that includes a visual field recovery trend map, pattern fitness analysis, and training suggestions.

Images