Eye Behavior Analysis for Mental State Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computer systems face challenges in accurately inferring human mental states from behavioral data, as facial expressions and vocal tones are under voluntary control and can be deceptive, while eye behaviors offer a more reliable and universally expressive indicator of mental states but have been understudied in machine learning applications.
Innovation Solution
A computational model using Bayesian deep belief networks to analyze eye behaviors such as pupil dilation, blink rate, and gaze movements to make real-time inferences about cognitive and emotional states, leveraging established computer vision and machine learning techniques to extract relevant features from video data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If facial expressions and vocal tones are used to infer mental states, then the system can capture a wide range of emotional indicators, but the reliability of inference decreases because these behaviors are under voluntary control and can be deceptive
Solution Approach 1:
The patent extracts and isolates eye behavior data from other facial expressions and vocal tones. By focusing specifically on eye behaviors (pupil dilation, blink rate, gaze movements) which are under involuntary control, the system removes the deceptive element from mental state inference while maintaining emotional indicator capability
Solution Approach 2:
The patent introduces eye behaviors as an intermediary indicator between voluntary expressions and actual mental states. Eye behaviors serve as a mediator that reflects genuine emotional and cognitive states without being subject to voluntary control, thus improving reliability while still providing emotional information
2Reliability
If eye behaviors are used to infer mental states, then the reliability of inference increases because eye behaviors are under involuntary control, but the amount of available research and machine learning applications is limited
Solution Approach 1:
The patent performs preliminary data collection and experimentation to establish the relationship between eye behaviors and mental states before deploying the machine learning model. By conducting controlled experiments to gather training data on pupil dilation, blink rate, and gaze movements under various cognitive and emotional conditions, the system prepares sufficient research data in advance to overcome the lack of existing studies
Solution Approach 2:
The patent creates a self-contained system that generates its own training data through controlled experiments and data collection. Rather than relying on existing external research, the system independently collects, processes, and utilizes eye behavior data to train its machine learning models, thus overcoming the limitation of unavailable research data
3Measurement precision
If multiple eye behavior features are analyzed simultaneously, then the accuracy of mental state inference improves, but the complexity of the computational model increases
Solution Approach 1:
The patent segments the analysis of eye behaviors into distinct features (pupil dilation, blink rate, blink duration, gaze movements) that can be processed separately and then integrated. This segmentation allows the computational model to handle complex multi-feature analysis by breaking it down into manageable components, thereby improving accuracy without overwhelming system complexity
Solution Approach 2:
The patent transitions from analyzing single eye behavior features to analyzing multiple features simultaneously across different dimensions. By incorporating temporal dynamics and multiple feature types together, the system achieves more accurate mental state inference while using machine learning techniques to manage the increased dimensionality
Data Source
AI summary
A method of discovering relationships between eye movements and emotional responses of a user starts with engaging the user in a plurality of tasks configured to elicit a predicted specific emotional response or allowing the user to experience a naturalistic environment with free-form social interactions. A first camera films at least one eye of the user recording a first time series of eye movements and a second camera films an outward looking view of the user recording a second time series of outward events perceived by the user. The first and second time series are sent to a computing device which compares the eye movements from the first time series and the outward events from the second time series. The computing device can then identify at least one relationship between eye movements that correlate to an actual specific emotional response.