Rapid Serial Visual Presentation Triage Using User State Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current rapid serial visual presentation (RSVP) systems for image triage are inefficient due to neglecting physical and cognitive states of users, leading to missed targets and inefficient processing of intelligence imagery, which limits the ability to extract useful information from vast datasets.
Innovation Solution
A method and system that prioritizes image triage output by monitoring user physical and cognitive states, using sensors like EEG, eye trackers, and gyroscopes to assign priorities based on optimal or suboptimal processing conditions, and employing real-time ERP detection techniques such as linear projections, nonlinear matched filters, and time-frequency distributions to enhance detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If RSVP presents images at high speed (tens or hundreds of milliseconds per image), then processing throughput increases, but user attention and processing effectiveness deteriorate
Solution Approach 1:
The system dynamically adjusts the image presentation rate based on real-time monitoring of user cognitive and physical states. When attention levels decline or cognitive load increases, the presentation rate is reduced to maintain processing effectiveness. This dynamic adaptation resolves the contradiction by allowing high throughput when users are attentive while preventing accuracy deterioration when users become fatigued or distracted.
2Extent of automation
If triage prioritization is based solely on ERP signals, then target detection automation increases, but false positives and false negatives increase due to unmonitored user states
Solution Approach 1:
The system incorporates continuous feedback from multiple sensors (eye trackers, EEG, gyroscopes) that monitor user attention, cognitive state, and physical stability. This feedback is integrated with ERP signal analysis to adjust triage prioritization in real-time. When users are in optimal states, ERP-based automation operates at high precision; when users deteriorate into suboptimal states, the system compensates by reducing automation reliance and flagging images for manual review, thereby maintaining measurement precision while preserving automation benefits.
3Measurement precision
If manual triage is performed by limited numbers of analysts, then processing accuracy can be maintained, but processing speed and coverage of vast imagery datasets deteriorate
Solution Approach 1:
The system introduces an intelligent intermediary layer that acts as a bridge between automated RSVP processing and human analyst review. This intermediary uses real-time user state monitoring to dynamically prioritize images, presenting only those most likely to contain targets to analysts. This resolves the contradiction by filtering out low-priority images that can be handled automatically, allowing analysts to focus their precision on high-value images while maintaining high overall processing throughput.
4Productivity
If images are processed during suboptimal user states, then processing volume increases, but the likelihood of missing targets increases
Solution Approach 1:
The system performs preliminary assessment of user cognitive and physical states before and during image processing. By monitoring attention levels, cognitive load, and physical stability in advance, the system can identify suboptimal states before they lead to missed targets. When suboptimal states are detected, the system proactively flags processed images for re-examination or manual review, ensuring that processing volume does not compromise target detection reliability.
Data Source
AI summary
A method and system are provided that prioritize the output of an image triage that is based on rapid serial visual presentation. User responses and estimates of the effectiveness with which each image is likely to have been processed by a user are employed for post triage image prioritization of potential targets. Images associated with a user response, processed during optimal user states, are assigned the highest priority for post triage examination, as targets are likely. Images without a user response that are processed during optimal user states are assigned the lowest priority, as these are unlikely to contain targets. Images with a user response that are processed during suboptimal states are assigned a medium priority, as these are likely to contain a high number of false positives. Images without a user response, processed during suboptimal user states are flagged for reprocessing as these may contain targets that the user may not have detected.