AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Learning from Human Gaze: Human-like Robot Social Navigation in Dense Crowds

Zhecheng Yu, Yan Lyu, Chen Yang, Tao Chen, Yishuang Zhang, Bo Ling, Peng Wang, Guanyu Gao, Weiwei Wu, Brian Y. Lim

AAAI 2026 · Vision-language and multimodal › Human and embodied modeling · approach: Human-like visual cognition in VLMs · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add gaze-based saliency filtering over all pedestrian masks (ablation)Gaze2Nav · GazeNav testADE: 0.2861 -> 0.2541Improved
add semantic saliency matching instead of raw gaze heatmaps (ablation)Gaze2Nav · GazeNav testADE: 0.2946 -> 0.2541Improved
add gaze regularization loss in Motion Planner (ablation)Gaze2Nav · GazeNav testADE: 0.2768 -> 0.2541Improved
add motion regularization loss in Gaze Predictor (ablation)Gaze2Nav · GazeNav testGaze MAE: 5.969 -> 3.876Improved
add motion regularization loss in Gaze Predictor (ablation)Gaze2Nav · GazeNav testADE: 0.2653 -> 0.2541Improved
predict gaze point first then match pedestrians, vs direct mask prediction (ablation)Gaze2Nav · GazeNav testSaliency F1: 0.5739 -> 0.8755Improved
add Gaze2Nav gaze pipeline vs best gaze baselineViNT+CGL · GazeNav testADE: 0.3005 -> 0.2541Improved
add CGL gaze alignment lossViNT · GazeNav testADE: 0.3160 -> 0.3005Improved
add Foveated gaze blurringGNM · GazeNav testADE: 0.4139 -> 0.3850Improved
Gaze2Nav vs ViNT in Overtaking scenarioViNT · GazeNav test, OvertakingADE: 0.3750 -> 0.2908Improved
Gaze2Nav vs ViNT in Following scenarioViNT · GazeNav test, FollowingADE: 0.2608 -> 0.2289Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy