Transformer model predicts driver's gaze object directly

Researchers have developed TransGaze-Object, a novel framework utilizing Transformer-based cross-attention to predict a driver's gaze object directly from facial and scene features. This approach bypasses the intermediate step of estimating a point-of-gaze, leading to more semantically meaningful attention representations. The framework achieves 60% accuracy in gaze-object prediction, a significant improvement over traditional point-of-gaze association methods, and demonstrates a substantial reduction in error rates. AI

IMPACT This research could enhance driver monitoring systems and autonomous vehicle perception by providing more accurate and semantically rich insights into driver attention.

RANK_REASON The cluster contains a research paper detailing a new model and dataset. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• arXiv
• Hugging Face
• Pavan Kumar Sharma
• Transformer
• TransGaze-Object
• Urban Driving-Face Scene Gaze

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →