EEGazeNet uses EEG to refine visual-search attention prediction
Original title: EEGaze: Multimodal Attention Prediction in Visual Search via EEG-Based Saliency Refinement
The public EEGaze 140 set pairs EEG and eye tracking from 36 people across 140 realistic scenes.
Researchers release EEGaze 140, described as the largest public multimodal dataset of free visual search with realistic stimuli. It contains synchronized EEG and eye-tracking recordings from 36 participants over 140 naturalistic scenes covering desktop-icon search and finding tools in cluttered workshops. EEGazeNet adds an EEG-guided stage that adapts a state-of-the-art saliency predictor to the user's cognitive state and, on EEGaze 140, outperforms all baselines and ablations.
- Prior computational attention models largely ignored EEG, the authors say.
- The refinement stage adapts an existing saliency model rather than replacing it.
- Two scenarios are desktop icons and tools in workshop scenes.
From the original article
1 min read at arxiv.orgVisual search is ubiquitous in daily life, from locating keys in a messy apartment to finding a tool in a cluttered workshop or an application icon on a desktop screen. While visual attention is to a large extent driven by image content, research in psychology indicates that user cognitive states have a significant influence on attention allocation.
By Mansi Sharma · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.