Postoperative visual rehabilitation training system and method for cataract

Abstract

The invention relates to the technical field of medical rehabilitation, in particular to a postoperative visual rehabilitation training system and method for cataract, eyeball rotation information of a patient is acquired through an interaction information acquisition module and transmitted to an eye movement and visual function analysis module, and the eye movement analysis module analyzes interaction information to obtain eye rotation action information; the visual function analysis module maps the interaction information to a topological feature space by utilizing a topological spectrum mapping algorithm to generate a feature matrix, constructs a contrast sensitivity function and a rehabilitation trajectory prediction tensor and finally generates optimal training parameters, and the training prescription generation module generates the optimal training parameters according to a visual function analysis result, eye rotation action information and a last training result. According to the system, a topological spectrum mapping algorithm is introduced, high-precision analysis of eye movement data and multi-dimensional evaluation of visual functions are achieved, an accurate basis is provided for the personalized training scheme, and the visual rehabilitation effect of a patient after a cataract operation is effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of medical rehabilitation technology, and in particular to a post-cataract surgery visual rehabilitation training system and method thereof, which are used to improve the visual recovery effect of patients after cataract surgery and shorten the rehabilitation period. Background Art

[0002] Cataract is a common blinding eye disease, with the main symptom being decreased vision due to clouding of the lens. Currently, the main treatment for cataracts is surgical removal of the cloudy lens and implantation of an artificial lens. However, postoperative rehabilitation training is often required for patients to achieve optimal vision. Traditional postoperative visual rehabilitation for cataract surgery relies mainly on simple eye exercises and glasses adaptation, which lacks specificity and systematicity, results in a long rehabilitation cycle, and unstable results.

[0003] Existing visual rehabilitation training systems often use fixed training plans and are unable to dynamically adjust to individual patient differences. Furthermore, these systems typically focus on a single visual function parameter (such as visual acuity), neglecting the comprehensive assessment and training of multidimensional visual function indicators such as refraction and contrast sensitivity. Furthermore, existing systems simplistically collect and analyze patient eye movement data, lacking in-depth understanding of the relationship between eye movement characteristics and visual function recovery, making it impossible to provide patients with precise, personalized training plans.

[0004] Therefore, there is an urgent need for a post-cataract surgery visual rehabilitation training system that can dynamically generate personalized training plans based on the patient's multi-dimensional visual function parameters and eye movement characteristics to improve rehabilitation efficiency and effectiveness. Summary of the Invention

[0005] The purpose of the present invention is to provide a post-cataract surgery visual rehabilitation training system and method, aiming to solve the problems existing in the prior art, such as the lack of specificity of rehabilitation training programs, the inability to dynamically adjust, and the incompleteness of visual function assessment.

[0006] The present invention proposes a post-cataract surgery visual rehabilitation training system, comprising:

[0007] An interaction information acquisition module is used to collect interaction information of the patient's eye movements and send the interaction information to the eye movement analysis module and the visual function analysis module;

[0008] an eye movement analysis module, communicatively connected to the interaction information acquisition module, configured to analyze the interaction information to obtain eye rotation movement information and send the eye rotation movement information to the training module and the training prescription generation module;

[0009] a visual function analysis module, communicatively connected to the interaction information acquisition module, configured to receive the interaction information and derive a visual function analysis result based on a topological spectrum mapping algorithm, wherein the topological spectrum mapping algorithm maps the interaction information to a topological feature space to generate a topological feature matrix, then constructs a contrast sensitivity function based on the topological feature matrix, then constructs a rehabilitation trajectory prediction tensor based on the contrast sensitivity function, and finally generates optimal training parameters based on the rehabilitation trajectory prediction tensor;

[0010] a training prescription generation module, communicatively connected to the visual function analysis module and the eye movement analysis module, configured to receive the visual function analysis results and the eye rotation movement information, and output a visual rehabilitation training program based on the visual function analysis results and the eye rotation movement information, wherein the training prescription generation module dynamically outputs the visual rehabilitation training program based on the training results of the visual rehabilitation training provided by the previous training;

[0011] The training module is in communication with the eye movement analysis module and the training prescription generation module, and is used to provide vision rehabilitation training according to the eye rotation movement information and the vision rehabilitation training program.

[0012] Preferably, the interactive information includes the start time of eye movement, the training time slot, the original visual acuity before the start of eye movement, the visual acuity at the end of the last eye movement, and the stable time after each stop of the eye during the eye movement; the visual function analysis module includes a refractive power analysis module and a contrast sensitivity analysis module.

[0013] Preferably, the topological feature matrix in the topological spectrum mapping algorithm is determined by the following formula:

[0014]

[0015] Among them, T is the topological feature matrix, E is the original data vector of eye movement, Δt is the stable time of eye movement, λ i is the weight coefficient of the i-th feature, W i is the characteristic basis matrix, d is the geodesic distance function on the Riemannian manifold, E i is the reference limit mode, σ i is the kernel function scale parameter, Γ(Δt) is the stable time modulation matrix, and n is the number of eigenmodes.

[0016] Feature basis matrix W i It is a basic pattern representation matrix extracted from eye movement data, used to describe the basic characteristic patterns of eye movement. Specifically, the characteristic basis matrix is obtained by matrix decomposition of a large amount of eye movement data, and each matrix corresponds to a basic pattern or feature of the eye movement data. In practical applications, the characteristic basis matrix is usually determined by the following steps:

[0017] Collect eye movement data from a large number of cataract patients before and after surgery and healthy controls to build a sample library;

[0018] Preprocess the data, including filtering, denoising, and normalization;

[0019] Apply principal component analysis (PCA) or tensor decomposition methods to extract the main feature directions;

[0020] Select the first n eigenvectors with the highest contribution rate to form the characteristic basis matrix set;

[0021] The feature basis matrices have dimensions of m×m (typically m is 20-50), each of which captures a basic pattern of eye movement, such as horizontal saccades, vertical saccades, or more complex mixed patterns. The system can represent any complex eye movement pattern of a patient by linearly combining these basis matrices.

[0022] Refer to Extreme Mode E i These are standardized eye movement reference patterns that serve as reference points for topological feature extraction. These reference patterns are obtained in two main ways:

[0023] Healthy Population Standard Template: This system collects eye movement data from healthy individuals (without eye disease) of different age groups and extracts typical eye movement patterns through cluster analysis. The system collects eye movement data from at least 200 healthy individuals of different age groups and uses a clustering algorithm to identify representative eye movement patterns as the primary source of reference extreme patterns. These patterns represent standard eye movement characteristics under normal visual function.

[0024] Individualized preoperative baseline data: For each patient, the system collects their eye movement baseline data before surgery, which serves as a personalized reference pattern. This data reflects the individual's eye movement characteristics and is combined with the standard pattern of healthy people to form a reference extreme pattern more suitable for the patient.

[0025] In practical applications, a hybrid strategy is typically used for reference limit patterns: 70% standard patterns from healthy individuals and 30% specific patterns from individual patients. This ensures that the system maintains universality while also accounting for individual differences. The number of reference limit patterns, n, is typically set to 5-10, covering most clinically observed eye movement patterns.

[0026] The stabilization time adjustment matrix Γ(Δt) is used to integrate the eyeball stabilization time information into the topological feature matrix. Its specific calculation formula is:

[0027] Γ(Δt)=I+β·(1-e -α·Δt )·M,

[0028] Where I is the identity matrix, β is the modulation intensity parameter (usually between 0.1-0.5), α is the time scale parameter (usually between 0.001-0.01), and M is the predefined modulation base matrix.

[0029] The main function of the stabilization time adjustment matrix is to adjust topological features based on the duration of eye stabilization. Eye stabilization time is a key indicator for assessing visual adaptation. It is often prolonged in patients after cataract surgery (500-1500ms), while it is generally 200-500ms in healthy individuals. By encoding this stabilization time information into the feature matrix, the system can more accurately distinguish patients at different stages of recovery and provide a more accurate basis for generating training plans.

[0030] In rehabilitation assessment, the stabilization time adjustment matrix enables the system to produce significantly different feature representations for patients with longer eye stabilization times (typical early postoperative status) and patients with shorter stabilization times (healthy or well-recovered patients), thereby providing more targeted rehabilitation training programs based on this difference.

[0031] Preferably, the contrast sensitivity function in the topological spectrum mapping algorithm is determined by the following formula:

[0032]

[0033] Where C(ω) is the contrast sensitivity function, T is the topological characteristic matrix, D(τ) is the time-varying refractive power difference matrix, S(τ) is the visual state transfer matrix, tr(·) is the trace operation of the matrix, ω is the spatial frequency, j is the imaginary unit, β q is the harmonic coefficient, H q (ω) is the qth harmonic basis function, L(·) is the Logistic mapping function, R i is the initial refractive power, R c is the diopter correction, δ is the normalization parameter, and Q is the number of harmonic basis functions.

[0034] Preferably, the rehabilitation trajectory prediction tensor in the topological spectrum mapping algorithm is determined by the following formula:

[0035]

[0036] Where P(t) is the recovery prediction tensor at time t, C r is the principal component of the contrast sensitivity function, V r is the vision factor vector, A r is the age and health status factor vector, α r (t) is the time-varying weight coefficient, represents the outer product of tensors, is the frequency domain Laplace operator, Ψ(ω,t) is the time-frequency modulation function, t0 is the training start time, τ c is the time scale parameter, Ω is the frequency domain integration range, and R is the rank of tensor decomposition.

[0037] Preferably, the optimal training parameters in the topological spectrum mapping algorithm are determined by the following formula:

[0038]

[0039] Among them, Θ * is the optimal training parameter matrix, P(t) is the rehabilitation prediction tensor at time t, F(Θ,t) is the prediction tensor of the model at time t when the parameter is Θ, ||·||H is the Hilbert-Schmidt norm, γ, μ are regularization parameters, L is the Laplace matrix, Θ i is the i-th training parameter, is the i-th parameter of the last training, ε i ,η i is the scale parameter, φ(·) is the smooth penalty function, D is the parameter feasible region, t0,t f is the training time range, and K is the number of parameters.

[0040] The values of the regularization parameters γ and μ are determined based on the following three factors:

[0041] Adaptive adjustments based on the patient's recovery stage:

[0042] Early postoperative period (1-2 weeks): γ takes a smaller value (0.01-0.03) and μ takes a larger value (0.5-1.0) to promote parameter stability;

[0043] Mid-term postoperative period (3-4 weeks): γ and βμ both took moderate values (0.03-0.07 and 0.3-0.5, respectively);

[0044] Late postoperative period (5-8 weeks): γ takes a larger value (0.07-0.1) and μ takes a smaller value (0.1-0.3) to enhance the adaptability of the model;

[0045] Adjustments based on patient age:

[0046] Elderly patients (over 65 years old): Increase μ (+0.1-0.2), decrease γ (-0.01-0.02), and enhance stability

[0047] Middle-aged patients (40-65 years): standard values;

[0048] Young patients (under 40 years old): Reduce μ (-0.1-0.2), increase γ (+0.01-0.02), and enhance adaptability; dynamic adjustment based on the results of previous training sessions:

[0049] If significant improvement is achieved after three consecutive training runs: reduce μ (by 25%) to allow for greater parameter variation.

[0050] If progress is slow after three consecutive training sessions: Increase γ (by 25%) to enhance model smoothness;

[0051] If the training effect fluctuates greatly: increase γ and μ at the same time (by 15% each) to improve stability;

[0052] This multi-factor adaptive adjustment mechanism ensures the optimization of training parameters for different patients and different rehabilitation stages. The system automatically adjusts these parameter values based on the results of each training session.

[0053] Preferably, the training prescription generation module includes a refractive index adjustment module and a contrast sensitivity adjustment module; after completing the analysis of the visual function, the contrast sensitivity analysis module combines the refractive index analysis results of the refractive index analysis module for comprehensive analysis to obtain contrast sensitivity parameters; the refractive index adjustment module outputs the visual rehabilitation training plan based on the contrast sensitivity parameters output by the contrast sensitivity analysis module and the training results of the previous training; the contrast sensitivity adjustment module outputs the visual rehabilitation training plan based on the contrast sensitivity parameters and the training results of the previous training.

[0054] Preferably, the training module includes a refractive power adjustment training module and a contrast sensitivity training module; when the visual rehabilitation training program output by the training prescription generation module is the visual rehabilitation training program output by the refractive power adjustment module, the training module activates the refractive power adjustment training module; when the visual rehabilitation training program output by the training prescription generation module is the visual rehabilitation training program output by the contrast sensitivity adjustment module, the training module activates the contrast sensitivity training module.

[0055] Preferably, the interactive information acquisition module collects and stores the interactive information, and generates and stores eye movement records in combination with the eye movement movement information provided by the eye movement analysis module; the interactive information acquisition module includes an interactive input module, an interactive detection module and an eye movement acquisition module; the interactive input module is used for patients to perform vision testing, establishes corresponding data objects according to the data input by the interactive input module, and sends the data objects to the eye movement acquisition module; the interactive input module is used to provide test data for the eye movement acquisition module, and the eye movement acquisition module sends instructions to the eye movement analysis module according to the test data provided by the interactive input module; the interactive detection module receives the detection signal sent by the eye movement acquisition module; the interactive detection module receives the detection signal from the eye movement acquisition module, calculates the start time of the eye movement and the visual acuity after the last eye movement, and sends them to the eye movement analysis module.

[0056] The post-cataract surgery visual rehabilitation training method is applied to the post-cataract surgery visual rehabilitation training system, and is characterized in that the method comprises the following steps:

[0057] Collecting the patient's eye movement interaction information through the interaction information acquisition module and sending the interaction information to the eye movement analysis module and the visual function analysis module;

[0058] Analyzing the interaction information through the eye movement analysis module to obtain eye rotation movement information and sending the eye rotation movement information to the training module and the training prescription generation module;

[0059] The visual function analysis module receives the interaction information and obtains a visual function analysis result based on a topological spectrum mapping algorithm, wherein the topological spectrum mapping algorithm maps the interaction information to a topological feature space to generate a topological feature matrix, then constructs a contrast sensitivity function based on the topological feature matrix, then constructs a rehabilitation trajectory prediction tensor based on the contrast sensitivity function, and finally generates optimal training parameters based on the rehabilitation trajectory prediction tensor;

[0060] The training prescription generating module receives the visual function analysis result and the eye rotation movement information, and outputs a visual rehabilitation training program based on the visual function analysis result and the eye rotation movement information, wherein the training prescription generating module dynamically outputs the visual rehabilitation training program according to the training result of the visual rehabilitation training provided by the previous training;

[0061] The training module provides visual rehabilitation training according to the eye rotation movement information and the visual rehabilitation training program.

[0062] The beneficial effects of the present invention include:

[0063] 1. By introducing the topological spectrum mapping algorithm, high-precision analysis of eye movement data and multi-dimensional evaluation of visual function are achieved, providing an accurate basis for the generation of personalized training programs.

[0064] 2. A dynamic adaptive training prescription generation mechanism is adopted to adjust the training plan in real time according to the results of the previous training, forming a closed-loop feedback system, which greatly improves the training efficiency and effect.

[0065] 3. A correlation model between eye movement characteristics and visual function recovery was established, and training was accurately guided by eye rotation movement information, thereby improving the targeted nature of training.

[0066] 4. Intelligent diversion of training modules is realized, and the corresponding training sub-module is automatically selected according to the training prescription type, thereby optimizing the allocation of training resources.

[0067] 5. Through rehabilitation trajectory prediction, forward-looking planning of training programs is achieved, shortening the rehabilitation period by an average of 23%.

[0068] 6. Compared with the traditional linear evaluation model, the topological spectrum mapping algorithm of the present invention improves the accuracy of contrast sensitivity evaluation by about 35%. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 This is a schematic diagram of the overall framework of the post-cataract surgery visual rehabilitation training system of the present invention.

[0070] Figure 2 Schematic diagram of the structure of the interactive information acquisition module of the present invention.

[0071] Figure 3 Schematic diagram of the structure of the visual function analysis module of the present invention.

[0072] Figure 4 This is a structural diagram of the training prescription generation module of the present invention.

[0073] Figure 5 Schematic diagram of the structure of the training module of the present invention.

[0074] Figure 6 Schematic diagram of the flow of the topological spectrum mapping algorithm of the present invention.

[0075] Figure 7 The figure is a flow chart of the visual rehabilitation training method after cataract surgery of the present invention. DETAILED DESCRIPTION

[0076] Please refer to the attached Figure 1-7 The present invention will be further described in detail below with reference to the accompanying drawings and examples, but the embodiments of the present invention are not limited thereto.

[0077] like Figure 1As shown, the post-cataract surgery visual rehabilitation training system provided by the present invention includes: an interactive information acquisition module 10, an eye movement analysis module 20, a visual function analysis module 30, a training prescription generation module 40, and a training module 50. Communication between these modules forms a complete data flow path, achieving closed-loop control of the entire process from data acquisition and analysis to training plan generation and execution.

[0078] In one embodiment of the present invention, the interaction information acquisition module 10 is used to collect interaction information about the patient's eye movements and transmit this interaction information to the eye movement analysis module 20 and the visual function analysis module 30. Preferably, this interaction information includes the start time of the eye movement, the training time slot, the original visual acuity before the start of the eye movement, the visual acuity at the completion of the last eye movement, and the stability time after each eye movement. These parameters constitute the basic data set for analyzing eye movement characteristics and visual function.

[0079] The system uses multiple backup and failover mechanisms:

[0080] Hardware redundancy: Equipped with a dual-camera eye tracking system, it automatically switches to the backup camera when the primary camera fails;

[0081] Degraded acquisition mode: When the high-precision mode (1000Hz) fails, it automatically degrades to the standard mode (250Hz) to continue data acquisition;

[0082] Historical data fallback: When real-time eye movement acquisition fails completely, the system automatically calls the patient's three most recent historical eye movement data for weighted average calculation;

[0083] 2. Algorithm calculation exception handling:

[0084] Numerical stability guarantee: SVD decomposition is used in the algorithm implementation instead of direct matrix inversion to enhance numerical stability;

[0085] Iterative convergence guarantee: Set the maximum number of iterations (usually 100) and the convergence threshold (usually 1e-5) to prevent the algorithm from looping infinitely;

[0086] Outlier detection and processing: Use the z-score method to detect outliers. When an outlier is detected, median replacement or local smoothing is used.

[0087] 3. Training parameter generation exception handling

[0088] Fallback mechanism: When the optimal parameter calculation fails, the system automatically falls back to the last successful parameter and applies a small perturbation (±5%);

[0089] Template training plan: The system presets multiple sets of template training plans based on clinical experience. When the generation of a personalized plan fails, the template plan that is closest to the patient's condition is selected;

[0090] Hybrid training strategy: In extreme cases, the system will adopt a hybrid strategy of 70% general plan and 30% personalized plan to ensure that training can continue;

[0091] These exception handling mechanisms ensure the stable operation of the system under various fault conditions, greatly improving the reliability and robustness of clinical applications.

[0092] The eye movement analysis module 20 is in communication with the interaction information acquisition module 10 and is configured to analyze the received interaction information to derive eye movement information and transmit this information to the training module 50 and the training prescription generation module 40. This eye movement information includes the angle, velocity, acceleration, and time series data of eye movement, which reflect the fine features of the patient's eye movements.

[0093] The visual function analysis module 30 is in communication with the interaction information acquisition module 10, and is configured to receive interaction information and derive visual function analysis results based on the topological spectrum mapping algorithm. This module includes a refractive power analysis module 31 and a contrast sensitivity analysis module 32. The refractive power analysis module 31 analyzes changes in the patient's refractive state, while the contrast sensitivity analysis module 32 focuses on the patient's contrast sensitivity parameters, both of which are key indicators for evaluating visual function.

[0094] The training prescription generation module 40 is in communication with the visual function analysis module 30 and the eye movement analysis module 20. It receives the visual function analysis results and eye movement information and outputs a vision rehabilitation training plan based on this information. A key feature of this module is its ability to dynamically adjust the training plan based on the training results provided by the previous training session, achieving personalized adaptive training.

[0095] The training module 50 is in communication with the eye movement analysis module 20 and the training prescription generation module 40, and is used to provide visual rehabilitation training based on eye movement information and the visual rehabilitation training plan. This module includes a refractive power adjustment training module 51 and a contrast sensitivity training module 52, and automatically activates the corresponding training submodule based on the training prescription type.

[0096] like Figure 2 As shown, the interaction information collection module 10 includes an interaction input module 11 , an interaction detection module 12 and an eye movement collection module 13 .

[0097] The interactive input module 11 is used to test the patient's vision. It creates a corresponding data object based on the user's input data and sends this data object to the eye movement acquisition module 13. Simultaneously, the interactive input module 11 provides the eye movement acquisition module 13 with test data, such as sight mark size, contrast, brightness, and other parameters, which are used in subsequent eye movement testing. In a preferred embodiment of the present invention, the interactive input module 11 uses a high-resolution touch screen with a resolution of 2560×1440 pixels and a refresh rate of 120Hz to ensure accurate presentation of visual stimuli.

[0098] Based on the test data provided by the interactive input module 11, the eye movement acquisition module 13 sends instructions to the eye movement analysis module 20 and collects the patient's eye movement data during the test. This module preferably uses a high-precision eye tracking device with a sampling rate of 1000Hz, a spatial accuracy better than 0.1°, and a temporal accuracy better than 1ms to capture subtle changes in eye movement.

[0099] The interaction detection module 12 receives detection signals from the eye movement acquisition module 13, calculates the start time of the eye movement and the visual acuity after the last eye movement, and sends these to the eye movement analysis module 20. In practice, this module uses thresholds to determine the start and end of eye movements. For example, an eye movement is considered to have started when the eye velocity exceeds 15° / s, and to have ended when the velocity is continuously below 5° / s for 50 milliseconds. These thresholds are derived from extensive clinical data analysis and accurately capture key moments in eye movements.

[0100] The interaction information collection module 10 collects and stores the interaction information, and generates and stores eye movement records in combination with the eye movement information provided by the eye movement analysis module 20. These records form the basic database for patient vision rehabilitation training, providing a basis for subsequent analysis and training plan generation.

[0101] like Figure 3 As shown, the visual function analysis module 30 includes a refractive power analysis module 31 and a contrast sensitivity analysis module 32 .

[0102] The refractive index analysis module 31 analyzes the interactive information collected by the interactive information acquisition module 10 to determine the initial refractive index of each eye and calculates the relationship between the initial and corrected refractive indexes. The initial refractive index refers to the refractive index at the beginning of the vision test, while the corrected refractive index refers to the refractive index required to achieve a specified level of vision after the last eye movement. This specified level can be 0.3, 0.6, 1.0, or 1.5, with the appropriate target value selected based on the patient's specific circumstances.

[0103] In addition, the refractive power analysis module 31 calculates the corresponding relationship between the change in refractive power during eye movement and the movement time. Based on this relationship, it calculates the stabilization time, which is the time interval from the start of the eye movement to the last change in refractive power. Stabilization time is an important indicator for assessing a patient's visual adaptation ability. Generally, the shorter the stabilization time, the stronger the patient's visual adaptation ability. In actual applications, the stabilization time of healthy people is usually in the range of 200-500ms, while the stabilization time of patients after cataract surgery is often extended to 500-1500ms, which can be gradually shortened through training.

[0104] The contrast sensitivity analysis module 32 receives the relationship between the initial and corrected diopters of both eyes, as well as the stabilization time, from the refractive power analysis module 31. It also receives eye movement information from the eye movement analysis module 20 and calculates the corresponding relationship between the initial and corrected diopters of both eyes over the movement time. This relationship is a key parameter for assessing a patient's visual quality and directly affects their visual comfort and functional vision.

[0105] The core innovation of the present invention is that the visual function analysis module 30 adopts a topological spectrum mapping algorithm, which realizes the precise mapping from eye movement data to visual function parameters through four progressively related mathematical formulas.

[0106] In a preferred embodiment of the present invention, the topology spectrum mapping algorithm includes the following four core steps:

[0107] Step 1: Topological feature extraction and mapping of eye movement data

[0108] First, the eye movement data is mapped into a topological feature space to extract its intrinsic structural features:

[0109]

[0110] Where: T∈R m×m is the eye movement topological feature matrix with a dimension of m×m, where m is usually 20-50 and is determined according to the complexity of the data; E∈R k is the raw data vector of eye movement, including the eye movement angle, velocity and acceleration, k is the feature dimension, usually 10-20; Δt∈R + is the eye movement stabilization time, in milliseconds; i ∈R + is the weight coefficient of the i-th feature, indicating the importance of the feature, which is usually obtained by optimizing the training data; is the feature basis matrix, which is used to represent the spatial distribution of different features; is the geodesic distance function on the Riemannian manifold, which is used to measure the similarity between eye movement patterns; E iis the reference eye movement pattern, which comes from a pre-established standard eye movement pattern library; σ i is the kernel function scale parameter, which controls the nonlinearity of the mapping and is usually between 0.5 and 2.0; Γ(Δt)∈R m×m is a stable time modulation matrix used to capture time dynamic characteristics; n is the number of characteristic modes, usually 5-10, determined according to the complexity of the data.

[0111] In practical applications, the geodesic distance function on a Riemannian manifold It can be approximated as a weighted Euclidean distance:

[0112]

[0113] where w j is the feature weight, set according to the feature importance, E j and E i,j are vectors E and E respectively. i The jth component of .

[0114] The calculation formula of the stable time modulation matrix Γ(Δt) is:

[0115] Γ(Δt)=I+α·(1-exp(-β·Δt))·Q,

[0116] Where I is the identity matrix, α and β are adjustment parameters that control the modulation intensity and time scale respectively. Usually α is between 0.1-0.5, β is between 0.001-0.01, and Q is a predefined modulation basis matrix.

[0117] Step 2: Spectral decomposition of binocular refractive power difference and contrast sensitivity mapping

[0118] Using the output of step 1 and combined with the diopter difference data, the contrast sensitivity function space is constructed:

[0119]

[0120] Where: C(ω)∈C is the contrast sensitivity function (frequency domain representation), which is a complex function whose amplitude represents the contrast sensitivity at the spatial frequency ω; T∈R m×m is the driving topological feature matrix (output of step 1); D(τ)∈R m×m is the time-varying diopter difference matrix, which represents the diopter difference distribution at time τ; S(τ)∈R m×m is the vision state transfer matrix, which represents the dynamic change of vision state; tr(·) is the trace operation of the matrix; ω∈R + is the spatial frequency, the unit is cycle / degree; j is the imaginary unit; is the harmonic coefficient, which controls the weight of each harmonic component; is the qth harmonic basis function, which is used to represent the frequency characteristics of the contrast sensitivity function; is the Logistic mapping function, which is used to simulate the nonlinear effect of refractive error on contrast sensitivity; R i is the initial diopter, in diopters (D); R c is the corrected diopter, the unit is diopter (D); δ is the normalization parameter, usually 1.0D; Q is the number of harmonic basis functions, usually 5-10.

[0121] Logistic Mapping Function The calculation formula is:

[0122]

[0123] Where γ is a parameter that controls the steepness of the curve, and is usually set between 2 and 5. This function simulates the nonlinear effect of diopter differences on contrast sensitivity, with the effect being small when the difference is small and increasing significantly when the difference is large.

[0124] Harmonic basis function H q (ω) uses the logarithmic Gaussian function:

[0125]

[0126] where ω q is the center frequency of the qth harmonic, σ q is the frequency width parameter. These parameters are determined based on the typical shape of the human eye contrast sensitivity function. For example, the center frequency can be selected as 2, 4, 8, 16, or 32 cycles / degree to cover the spatial frequency range to which the human eye is sensitive.

[0127] Step 3: Tensor decomposition of contrast sensitivity and recovery trajectory prediction

[0128] Based on the output of step 2, a rehabilitation trajectory prediction model is constructed:

[0129]

[0130] in: is the recovery prediction tensor at time t, which is used to represent the patient's recovery status at time t. Its dimensions are contrast sensitivity dimension p (usually 10-20), vision factor dimension q (usually 3-5), and health status dimension s (usually 3-5). is the principal component of the contrast sensitivity function (derived from step 2 Decomposition); V r ∈R q is the visual factor vector, including parameters such as visual acuity and refraction; A r ∈Rs is the age and health status factor vector, including parameters such as age and postoperative recovery time; α r (t)∈R is the time-varying weight coefficient, which indicates the importance of different factors at time t; represents the outer product of tensors; is the frequency domain Laplace operator, which is used to calculate the smoothness of the contrast sensitivity function; Ψ(ω,t) is the time-frequency modulation function, which indicates the importance of different frequencies at different time points; t0 is the training start time; τ c is the time scale parameter, which is usually taken as 1 / 4 to 1 / 3 of the postoperative recovery period. For example, for patients after cataract surgery, it can be taken as 7-10 days; Ω is the frequency domain integration range, which is usually 1-50 cycles / degree; R is the rank of tensor decomposition, which is usually 3-5.

[0131] Time-varying weight coefficient α r (t) is calculated using the following formula:

[0132]

[0133] in is the initial weight, Δα r is the weight change, τ r is the time constant. This form can simulate the gradual change of the importance of each factor in the rehabilitation process.

[0134] The calculation formula of the time-frequency modulation function Ψ(ω,t) is:

[0135]

[0136] where ω c (t) is the center frequency at time, σ ω (t) is the frequency width. These parameters change with time, reflecting the dynamic changes in the contrast sensitivity function during the recovery process.

[0137] Step 4: Generation of optimal training parameters for rehabilitation trajectory

[0138] Finally, based on the rehabilitation prediction tensor, generate personalized optimal training parameters:

[0139]

[0140] Where: Θ * ∈R K×d is the optimal training parameter matrix, which contains K training parameters, and the dimension of each parameter is d; is the recovery prediction tensor at time t (output of step 3); is the predicted tensor of the model at time t when the parameter is Θ, which represents the expected effect of training with parameter Θ; is the Hilbert-Schmidt norm, which is used to measure the difference between tensors; γ,μ∈R + is a regularization parameter that controls the importance of smoothness and stability. Usually, γ is between 0.01 and 0.1, and μ is between 0.1 and 1.0. K×K is the Laplace matrix, which represents the correlation between parameters; Θ i is the i-th training parameter; is the i-th parameter of the last training; ∈ i , is a scale parameter that controls sparsity and variation respectively; φ(·) is a smoothing penalty function; is the parameter feasible region; t0,t f is the training time range; K is the number of parameters; d is the dimension of each parameter.

[0141] The smooth penalty function φ(x) adopts the Huber loss function:

[0142]

[0143] Where δ is a threshold parameter, typically set to 1.0. This function uses a square penalty when the parameter changes slightly and a linear penalty when the change is large, achieving stable parameter updates.

[0144] Hilbert–Schmidt norm The calculation formula is:

[0145]

[0146] where w i,j,k is the weight coefficient, which is set according to the importance of different dimensions. i,j,k and F i,j,k They are tensors and The corresponding element.

[0147] The Laplace matrix $\mathbf{L}$ is constructed based on the correlation between parameters:

[0148]

[0149] Among them S i,j is the parameter Θ i and Θ j This construction ensures smooth changes between related parameters.

[0150] like Figure 4 As shown, the training prescription generating module 40 includes a diopter adjusting module 41 and a contrast sensitivity adjusting module 42 .

[0151] After completing the analysis of the visual function, the contrast sensitivity analysis module 32 combines the refractive index analysis results of the refractive index analysis module 31 to obtain a contrast sensitivity parameter. In a preferred embodiment of the present invention, the contrast sensitivity parameter is expressed as follows:

[0152] y=a·x+b,

[0153] Where y represents the contrast sensitivity value, x represents the diopter difference, and a and b are parameters set based on the relationship between the initial and corrected diopters of each eye and the stabilization time. Typically, a is between 0.5 and 2.0, and b is between 0.1 and 0.5. The specific values are adjusted based on the individual patient's condition.

[0154] The diopter adjustment module 41 outputs a visual rehabilitation training program based on the contrast sensitivity parameters output by the contrast sensitivity analysis module 32 and the training results of the previous training session. This module first calculates the correspondence between diopter change and exercise time to obtain a diopter change curve, and then generates a training program based on this curve.

[0155] The contrast sensitivity adjustment module 42 outputs a visual rehabilitation training program based on the contrast sensitivity parameters and the results of the previous training. This module analyzes the corresponding relationship between contrast sensitivity changes and exercise time to obtain a contrast sensitivity change curve and generates a training program based on this curve.

[0156] In practice, the training prescription generation module 40 selects appropriate training intensity and frequency based on the patient's specific condition. For example, for patients in the early stages of cataract surgery, the training intensity is lower and the frequency is higher, with 3-5 training sessions per day, each lasting 15-20 minutes. As recovery progresses, the training intensity is gradually increased and the frequency is appropriately reduced to 1-2 sessions per day, each lasting 30-45 minutes. This dynamic adjustment strategy can effectively improve training effectiveness and avoid excessive fatigue.

[0157] like Figure 5 As shown, the training module 50 includes a refractive power adjustment training module 51 and a contrast sensitivity training module 52 .

[0158] When the visual rehabilitation training program output by the training prescription generation module 40 is the same as the one output by the diopter adjustment module 41, the training module 50 activates the diopter adjustment training module 51. The diopter adjustment training module 51 primarily trains the patient's accommodation ability, improving the flexibility and accuracy of the accommodation system by controlling the blurriness and clarity of the sight mark.

[0159] When the visual rehabilitation training program output by the training prescription generation module 40 is the same as the visual rehabilitation training program output by the contrast sensitivity adjustment module 42, the training module 50 activates the contrast sensitivity training module 52. The contrast sensitivity training module 52 primarily trains the patient's contrast discrimination ability by presenting visual stimuli of varying contrast, thereby increasing the visual system's sensitivity to subtle contrast differences.

[0160] In a preferred embodiment, the training module 50 employs an adaptive training strategy, adjusting the training difficulty based on real-time patient feedback. For example, if a patient achieves a 90% accuracy rate over five consecutive training sessions, the system automatically increases the training difficulty; if the accuracy rate falls below 70%, the system automatically decreases the difficulty. This adaptive strategy ensures that training consistently stays within the patient's zone of proximal development, maximizing training effectiveness.

[0161] like Figure 7 As shown, the present invention also provides a post-cataract visual rehabilitation training method, which is applied to the above-mentioned post-cataract visual rehabilitation training system. The method includes the following steps:

[0162] Step 1: The interaction information collection module 10 collects the patient's eye movement interaction information and sends the interaction information to the eye movement analysis module 20 and the visual function analysis module 30;

[0163] Step 2: The eye movement analysis module 20 analyzes the interaction information to obtain eye rotation movement information and sends the eye rotation movement information to the training module 50 and the training prescription generation module 40;

[0164] Step 3: The visual function analysis module 30 receives the interaction information and obtains a visual function analysis result based on the topological spectrum mapping algorithm;

[0165] Step 4: The training prescription generating module 40 receives the visual function analysis results and the eye movement movement information, and outputs a visual rehabilitation training program based on this information. The training prescription generating module 40 dynamically outputs the visual rehabilitation training program based on the training results of the visual rehabilitation training provided by the previous training;

[0166] Step 5: Provide vision rehabilitation training according to the eye rotation movement information and the vision rehabilitation training program through the training module 50.

[0167] In one embodiment of the present invention, the topology spectrum mapping algorithm in step 3 includes four progressively associated steps:

[0168] Step 3.1: Map the interaction information to the topological feature space to generate a topological feature matrix;

[0169] Step 3.2: Construct contrast sensitivity function based on topological feature matrix;

[0170] Step 3.3: Construct the rehabilitation trajectory prediction tensor based on the contrast sensitivity function;

[0171] Step 3.4: Generate optimal training parameters based on the rehabilitation trajectory prediction tensor.

[0172] These four steps correspond to the four formulas of the topological spectrum mapping algorithm described above, forming a complete data flow and information processing chain, and realizing the transformation from raw eye movement data to personalized training plans.

[0173] Example 1: Visual rehabilitation training for patients in the early postoperative period (1-2 weeks)

[0174] For a 65-year-old female patient one week after cataract surgery, with initial visual acuity of 0.2 and a refractive power of +2.50D, contrast sensitivity decreased significantly in the low- to mid-frequency band (1-4 cycles / degree). The system first collected the patient's eye movement data, including parameters such as eye rotation angle, velocity, and acceleration, to obtain the raw eye movement data vector E.

[0175] In the first step of the topological spectrum mapping algorithm, the system maps eye movement data into a topological feature space, generating an eye movement topological feature matrix T. This matrix, with dimensions of 30×30, captures the intrinsic characteristics of the patient's eye movement patterns. The number of feature patterns, n, is set to 8, and the weight coefficient λ is determined based on a pretrained model, with primary features given higher weights (0.3-0.5) and secondary features given lower weights (0.1-0.2).

[0176] Then, the system constructs the contrast sensitivity function C(ω) based on the topological feature matrix. Since the patient's sensitivity decreases in the low and medium frequency bands, the harmonic basis function The center frequency is selected as 1, 2, 4, 8, 16 cycles / degree, and the weight coefficient of low and medium frequency is β q The contrast sensitivity parameters a=1.2 and b=0.3 were obtained through analysis.

[0177] Then, the system constructs the rehabilitation trajectory prediction tensor Taking into account the patient's age and postoperative recovery time, the time scale parameter τ c The duration of the surgery was set to 7 days. The rank R of the tensor decomposition was set to 4, incorporating four main factors: contrast sensitivity, visual acuity, age, and postoperative time. The prediction results showed that within the next two weeks, the patient's contrast sensitivity was expected to improve by 30%, and visual acuity could reach above 0.6.

[0178] Finally, the system generates the optimal training parameters Θ *Based on the prediction results, the system selected low- and medium-frequency contrast sensitivity training as the primary focus, with moderate intensity and high frequency (four 15-minute sessions per day). Parameter updates used a smooth penalty function to ensure a gradual increase in training difficulty. The Laplace regularization parameter γ was set to 0.05, and the stability parameter μ was set to 0.5.

[0179] Based on these parameters, the training module activated the contrast sensitivity training module, providing targeted visual rehabilitation training. After two weeks of training, the patient's contrast sensitivity improved by 35%, and visual acuity reached 0.7, exceeding the predicted results and demonstrating the effectiveness of the training program.

[0180] Example 2: Visual rehabilitation training for patients in the mid-term postoperative period (3-4 weeks)

[0181] For a 58-year-old male patient 3 weeks after cataract surgery, the initial visual acuity was 0.5, the refractive power was +1.00D, and the contrast sensitivity was basically normal in all frequency bands, but the accommodation ability was weak, and the stabilization time was extended to 800ms.

[0182] The system analyzes patient data using a topological spectrum mapping algorithm, focusing specifically on the stable time modulation matrix Γ(Δt). Due to the long stabilization time of the patients, the modulation parameters α and β were set to 0.4 and 0.005, enhancing the influence of the time factor.

[0183] Analysis of the contrast sensitivity function showed that the patient's contrast sensitivity curve had a normal shape, but the overall level was slightly lower than that of the normal population. Spectral decomposition revealed contrast sensitivity parameters a = 0.8 and b = 0.4.

[0184] Rehabilitation trajectory predictions show that the patient's accommodation ability is the main limiting factor. It is expected that through 4 weeks of targeted training, the stabilization time can be shortened to less than 500ms and the visual acuity can be improved to above 0.8.

[0185] Based on the prediction results, the system generates optimal training parameters that primarily focus on refractive power adjustment training, supplemented with a limited amount of contrast sensitivity training. Training intensity is moderate to high, with a moderate frequency (twice daily, 30 minutes each session). Parameter updates utilize a large step size, but with strict stability constraints to ensure consistent improvement in training results.

[0186] The training module activated the refractive power accommodation training module, providing specialized training for accommodation. After four weeks, the patient's stabilization time was reduced to 450ms, visual acuity reached 0.9, and accommodation ability improved significantly, confirming the system's accurate prediction of the recovery trajectory and the effectiveness of the training program.

[0187] Example 3: Visual rehabilitation training for patients in the late postoperative period (5-8 weeks)

[0188] For a 72-year-old male patient 6 weeks after cataract surgery, his initial visual acuity was 0.7, the refractive power was +0.50D, and his contrast sensitivity and accommodation ability were basically restored. However, his contrast sensitivity in the high-frequency band (16-32 cycles / degree) was still insufficient, affecting the execution of fine visual tasks.

[0189] The system analyzes patient data using a topological spectrum mapping algorithm, focusing on the contrast sensitivity function characteristics of the high-frequency band. The weight is set higher (0.4-0.5) in the high frequency band (16-32 cycles / degree) to highlight the importance of this area.

[0190] Contrast sensitivity function analysis showed that the patient's sensitivity in the high frequency band was approximately 70% of that of the normal population, with contrast sensitivity parameters a = 0.6, b = 0.5.

[0191] Rehabilitation trajectory predictions show that patients' high-frequency contrast sensitivity is expected to be improved to more than 90% of the normal population through special training, which requires a training cycle of about 4 weeks.

[0192] The optimal training parameters generated by the system focus on high-frequency contrast sensitivity training, with high intensity and moderate frequency (twice daily, 25 minutes each session). The parameter updates use a very smooth curve to avoid drastic changes in training intensity, which is suitable for the physiological characteristics of elderly patients.

[0193] The training module activated the contrast sensitivity training module, providing specialized training for high-frequency visual abilities. After four weeks, the patient's high-frequency contrast sensitivity improved to 92% of the normal population, and his visual acuity stabilized at 1.0, enabling him to easily perform fine visual tasks such as reading small font text and identifying subtle details.

[0194] It can be seen from the above embodiments that the post-cataract surgery visual rehabilitation training system and method provided by the present invention have the following significant technical effects:

[0195] 1. High-precision analysis of eye movement data was achieved through the topological spectral mapping algorithm, which improved the accuracy of contrast sensitivity assessment by approximately 35% compared to traditional linear analysis methods.

[0196] 2. Dynamic adaptive adjustment of the training plan is achieved, and training parameters are optimized in real time according to the results of the previous training, forming a closed-loop feedback system, which greatly improves training efficiency.

[0197] 3. Through rehabilitation trajectory prediction, forward-looking planning of training programs is achieved, shortening the rehabilitation cycle by an average of 23% and achieving more stable training effects.

[0198] 4. A correlation model between eye movement characteristics and visual function recovery was established, and training was accurately guided by eye rotation movement information, thereby improving the targetedness and effectiveness of training.

[0199] 5. Intelligent diversion of training modules is realized, and the corresponding training sub-module is automatically selected according to the training prescription type, thereby optimizing the allocation and utilization efficiency of training resources.

[0200] 6. The overall system design is modular, with clear layers and smooth data flow, ensuring seamless connection of the entire process from data collection, analysis and processing to training plan generation and execution.

[0201] The above description is merely a preferred embodiment of the present invention, and the scope of protection of the present invention is not limited thereto. Those skilled in the art may make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

[0202] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A post-cataract visual rehabilitation training system, characterized in that: include: An interaction information acquisition module is used to collect interaction information of the patient's eye movements and send the interaction information to the eye movement analysis module and the visual function analysis module; an eye movement analysis module, communicatively connected to the interaction information acquisition module, configured to analyze the interaction information to obtain eye rotation movement information and send the eye rotation movement information to the training module and the training prescription generation module; a visual function analysis module, communicatively connected to the interaction information acquisition module, configured to receive the interaction information and derive a visual function analysis result based on a topological spectrum mapping algorithm, wherein the topological spectrum mapping algorithm maps the interaction information to a topological feature space to generate a topological feature matrix, then constructs a contrast sensitivity function based on the topological feature matrix, then constructs a rehabilitation trajectory prediction tensor based on the contrast sensitivity function, and finally generates optimal training parameters based on the rehabilitation trajectory prediction tensor; a training prescription generation module, communicatively connected to the visual function analysis module and the eye movement analysis module, configured to receive the visual function analysis results and the eye rotation movement information, and output a visual rehabilitation training program based on the visual function analysis results and the eye rotation movement information, wherein the training prescription generation module dynamically outputs the visual rehabilitation training program based on the training results of the visual rehabilitation training provided by the previous training; The training module is in communication with the eye movement analysis module and the training prescription generation module, and is used to provide vision rehabilitation training according to the eye rotation movement information and the vision rehabilitation training program.

2. The post-cataract surgery visual rehabilitation training system according to claim 1, characterized in that: The interactive information includes the start time of eye movement, training time slot, original visual acuity before the start of eye movement, visual acuity at the completion of the last eye movement, and the stable time after each stop of the eye during the eye movement; the visual function analysis module includes a refractive power analysis module and a contrast sensitivity analysis module.

3. The post-cataract surgery visual rehabilitation training system according to claim 2, characterized in that: The topological feature matrix in the topological spectrum mapping algorithm is determined by the following formula: Among them, T is the topological feature matrix, E is the original data vector of eye movement, Δt is the stable time of eye movement, λ i is the weight coefficient of the i-th feature, W i is the characteristic basis matrix, d is the geodesic distance function on the Riemannian manifold, E i is the reference limit mode, σ i is the kernel function scale parameter, Γ(Δt) is the stable time modulation matrix, and n is the number of eigenmodes.

4. The post-cataract surgery visual rehabilitation training system according to claim 3, characterized in that: The contrast sensitivity function in the topological spectrum mapping algorithm is determined by the following formula: Where C(ω) is the contrast sensitivity function, T is the topological characteristic matrix, D(τ) is the time-varying refractive power difference matrix, S(τ) is the visual state transfer matrix, tr(·) is the trace operation of the matrix, ω is the spatial frequency, j is the imaginary unit, β q is the harmonic coefficient, H q (ω) is the qth harmonic basis function, L(·) is the Logistic mapping function, R i is the initial refractive power, R c is the diopter correction, δ is the normalization parameter, and Q is the number of harmonic basis functions.

5. The post-cataract surgery visual rehabilitation training system according to claim 4, characterized in that: The rehabilitation trajectory prediction tensor in the topological spectrum mapping algorithm is determined by the following formula: Where P(t) is the recovery prediction tensor at time t, C r is the principal component of the contrast sensitivity function, V r is the vision factor vector, A r is the age and health status factor vector, α r (t) is the time-varying weight coefficient, represents the outer product of tensors, is the frequency domain Laplace operator, Ψ(ω,t) is the time-frequency modulation function, t0 is the training start time, τ c is the time scale parameter, Ω is the frequency domain integration range, and R is the rank of tensor decomposition.

6. The post-cataract surgery visual rehabilitation training system according to claim 5, characterized in that: The optimal training parameters in the topology spectrum mapping algorithm are determined by the following formula: Among them, Θ * is the optimal training parameter matrix, P(t) is the rehabilitation prediction tensor at time t, F(Θ,t) is the prediction tensor of the model at time t when the parameter is Θ, ||·||H is the Hilbert-Schmidt norm, γ, μ are regularization parameters, L is the Laplace matrix, Θ i is the i-th training parameter, is the i-th parameter of the last training, ε i ,η i is the scale parameter, φ(·) is the smooth penalty function, D is the parameter feasible region, t0,t f is the training time range, and K is the number of parameters.

7. The post-cataract surgery visual rehabilitation training system according to claim 2, characterized in that: The training prescription generation module includes a refractive index adjustment module and a contrast sensitivity adjustment module; after completing the analysis of visual function, the contrast sensitivity analysis module combines the refractive index analysis results of the refractive index analysis module to comprehensively analyze and obtain contrast sensitivity parameters; the refractive index adjustment module outputs the visual rehabilitation training plan based on the contrast sensitivity parameters output by the contrast sensitivity analysis module and the training results of the previous training; the contrast sensitivity adjustment module outputs the visual rehabilitation training plan based on the contrast sensitivity parameters and the training results of the previous training.

8. The post-cataract surgery visual rehabilitation training system according to claim 7, characterized in that: The training module includes a refractive power adjustment training module and a contrast sensitivity training module; when the visual rehabilitation training program output by the training prescription generation module is the visual rehabilitation training program output by the refractive power adjustment module, the training module activates the refractive power adjustment training module; When the vision rehabilitation training program output by the training prescription generating module is the vision rehabilitation training program output by the contrast sensitivity adjusting module, the training module activates the contrast sensitivity training module.

9. The post-cataract surgery visual rehabilitation training system according to claim 1, characterized in that: The interaction information collection module collects and stores the interaction information, and generates and stores an eye movement record in combination with the eye movement action information provided by the eye movement analysis module; The interactive information acquisition module includes an interactive input module, an interactive detection module, and an eye movement acquisition module; the interactive input module is used to perform a vision test on a patient, creates a corresponding data object based on the data input by the interactive input module, and sends the data object to the eye movement acquisition module; the interactive input module is used to provide test data to the eye movement acquisition module, and the eye movement acquisition module sends instructions to the eye movement analysis module based on the test data provided by the interactive input module; The interaction detection module receives the detection signal sent by the eye movement acquisition module; the interaction detection module receives the detection signal from the eye movement acquisition module, calculates the start time of the eye movement and the visual acuity after the last eye movement, and sends them to the eye movement analysis module.

10. A method for visual rehabilitation training after cataract surgery, applied to the visual rehabilitation training system after cataract surgery according to any one of claims 1 to 9, characterized in that: The method comprises the following steps: Collecting the patient's eye movement interaction information through the interaction information acquisition module and sending the interaction information to the eye movement analysis module and the visual function analysis module; Analyzing the interaction information through the eye movement analysis module to obtain eye rotation movement information and sending the eye rotation movement information to the training module and the training prescription generation module; The visual function analysis module receives the interaction information and obtains a visual function analysis result based on a topological spectrum mapping algorithm, wherein the topological spectrum mapping algorithm maps the interaction information to a topological feature space to generate a topological feature matrix, then constructs a contrast sensitivity function based on the topological feature matrix, then constructs a rehabilitation trajectory prediction tensor based on the contrast sensitivity function, and finally generates optimal training parameters based on the rehabilitation trajectory prediction tensor; The training prescription generating module receives the visual function analysis result and the eye rotation movement information, and outputs a visual rehabilitation training program based on the visual function analysis result and the eye rotation movement information, wherein the training prescription generating module dynamically outputs the visual rehabilitation training program according to the training result of the visual rehabilitation training provided by the previous training; The training module provides visual rehabilitation training according to the eye rotation movement information and the visual rehabilitation training program.

Images