Explainable AI for Computer Vision Syndrome Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computer Vision Syndrome (CVS) is prevalent among individuals who spend extended periods in front of screens, leading to eye strain and fatigue, with existing solutions failing to provide effective predictive and remedial measures, particularly in terms of trust and interpretability in AI-driven interventions.
Innovation Solution
A system utilizing machine learning and explainable artificial intelligence (XAI) to determine the potential development of CVS by analyzing input parameters such as blink rate, ambient light, and other environmental factors, and recommending remedial actions based on weighted parameter importance, ensuring transparency and user trust.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If AI algorithms are used to predict CVS, then prediction accuracy is improved, but interpretability and user trust deteriorate
Solution Approach 1:
The patent introduces explainable AI (XAI) as an intermediary layer between the black-box ML model and the user. The XAI component translates complex model predictions into human-understandable explanations, showing which input parameters (blink rate, ambient light, etc.) contributed most to the CVS prediction. This mediator preserves the high accuracy of the underlying ML model while providing transparency through feature importance visualization and decision breakdowns.
Solution Approach 2:
The patent changes the representation of model outputs from raw probability scores to normalized parameter contributions. By transforming the output into a format that shows the relative importance of each input parameter (e.g., 'blink rate contributed 40% to the prediction'), the system maintains prediction accuracy while improving interpretability. The parameter transformation makes the decision-making process transparent to users.
2Measurement precision
If multiple input parameters are analyzed, then prediction accuracy is improved, but system complexity increases
Solution Approach 1:
The patent segments the input parameters into distinct categories (eye movement parameters, environmental parameters, device parameters) and processes them through separate feature extraction modules. Each parameter type is handled by dedicated processing logic that normalizes and prepares the data independently, then feeds into the unified ML model. This segmentation reduces overall system complexity by making the data flow modular and manageable.
Solution Approach 2:
The patent implements a universal data processing pipeline that handles multiple input parameter types through a common architecture. The same feature extraction and normalization mechanisms work for all parameter categories (blink rate, ambient light, device distance, etc.), reducing redundancy. The unified model structure processes all parameters through consistent logic, simplifying the system compared to separate specialized handlers for each parameter type.
Data Source
AI summary
Technology for applying explainable artificial training algorithms (XAI) to training machine learning algorithms for identifying potentially developing computer vision syndrome (CVS), CVS and/or recommended remedial action(s) that a user can perform to counter potentially developing CVS and/or existing CVS. In some embodiments, the XAI includes a Contrastive Explainability model. In some embodiments, the training performed by the XAI includes assigning weight factors respectively to CVS input parameters (for example, blink rate) based upon how strong the respective CVS input factor is correlated with development of CVS in the user.