Visual Attention Model Accuracy Improvement via Behavioral Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current visual attention models are limited in accuracy and practicality for commercial applications, as they struggle to perfectly simulate human visual attention allocation and require invasive and expensive eye-tracking studies for validation.
Innovation Solution
A system and method for improving visual attention models by assessing the relative effectiveness of environments on influencing human behavior using data indicative of attention allocation, allowing for the modification of models based on collected data to enhance accuracy without the need for direct eye-tracking measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If eye-tracking studies are used to validate visual attention models, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent uses behavioral data (clicks, purchases, survey responses) as an intermediary indicator to infer visual attention allocation. Instead of directly measuring eye movements with complex eye-tracking equipment, the system measures easier-to-obtain behavioral outcomes that are correlated with attention, thereby resolving the contradiction between measurement precision and device complexity
Solution Approach 2:
The patent replaces the mechanical/optical eye-tracking system with computational models that process behavioral data. Instead of using physical eye-tracking devices to measure attention, the system uses software-based visual attention models that simulate human attention and are validated against behavioral outcomes, eliminating the need for complex hardware
2Reliability
If eye-tracking studies are conducted for model validation, then reliability is improved, but loss of time and resources increase
Solution Approach 1:
The patent uses readily available behavioral data (clicks, purchases, survey responses) as copies or proxies for direct attention measurements. These behavioral copies serve as validation metrics that are much faster and cheaper to obtain than conducting full eye-tracking studies, while still providing reliable validation of visual attention model predictions
Solution Approach 2:
The patent performs preliminary model validation using easily collected behavioral data before investing in comprehensive eye-tracking studies. By first testing model predictions against available behavioral metrics, the system can quickly identify promising models without time-consuming detailed eye-tracking analysis
3Measurement precision
If visual attention models are made more complex to improve accuracy, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent applies partial action by using simplified visual attention models that capture the most important attention-determining features without implementing all possible complexity. The models focus on key visual characteristics that drive attention allocation, achieving sufficient precision without unnecessary complexity that would hinder practical deployment
Data Source
AI summary
Systems and methods for improving visual attention models use effectiveness assessment from an environment as feedback to improve visual attention models. The effectiveness assessment uses data indicative of a particular behavior, which is related to visual attention allocation, received from the environment to assess relative effectiveness of the environment on influencing the particular behavior.