Detection and Differentiation of Obstacles in Repeated Adaptive Human-Computer Interactions from ...
summary
System implements a matching-card task with two induced obstacles: a memory load via a cumulative-sum secondary task and a visual impairment by showing brown-tinted cards; adaptations are either repeating previously seen cards or adding auditory identifiers for positions. Real-time detectors use EEG and interaction logs, fused by a dynamic Bayesian network to pick adaptations across sessions. Initial decision accuracy averages 72.5% and climbs to 98% after additional interactions, showing the multimodal fusion corrects earlier wrong adaptations.
AI-written summary · about
chapters (8)
- 0:00Introduction and motivation
- 1:47Obstacle detection techniques
- 3:37Multimodal data approach
- 5:26Game design and obstacles
- 7:30System architecture overview
- 8:16Evaluation and results
- 9:38Model correction analysis
- 10:05Conclusion and summary
related talks
06mDeriving Effective Human Activity Recognition Systems through Objective Task Complexity Assessment
05mLeveraging Activity Recognition to Enable Protective Behavior Detection in Continuous Data
06mA Systematic Study of Unsupervised Domain Adaptation for Robust Human-Activity Recognition
39mMind-Reading with Adaptive and Intelligent User Interfaces
36mMachine-Brain Intelligence in Brain-Computer Interfaces
13mBeyond Barriers: AI Reshaping Digital Accessibility
01h 05mAdaptive Generalization: The Role of Dynamic Evaluation in Low-Resource Settings
13mEnhancing Object Detection & Avoidance Systems for Robots in Challenging Environments Using AI