Search papers, labs, and topics across Lattice.
This paper introduces an end-to-end method for shared attention (SA) estimation that simultaneously detects groups of people and estimates the SA heatmap. The approach uses a two-step process involving initial SA heatmap generation based on individual gaze and group membership, followed by a refinement of group memberships using the initial SA heatmap. Experiments demonstrate improved performance in both group detection and SA estimation compared to existing methods that either don't detect groups or assume a single SA point.
Detecting who is paying attention is now more accurate, thanks to a new method that jointly infers group membership and shared attention.
This paper proposes an end-to-end shared attention estimation method via group detection. Most previous methods estimate shared attention (SA) without detecting the actual group of people focusing on it, or assume that there is a single SA point in a given image. These issues limit the applicability of SA detection in practice and impact performance. To address them, we propose to simultaneously achieve group detection and shared attention estimation using a two step process: (i) the generation of SA heatmaps relying on individual gaze attention heatmaps and group membership scalars estimated in a group inference; (ii) a refinement of the initial group memberships allowing to account for the initial SA heatmaps, and the final prediction of the SA heatmap. Experiments demonstrate that our method outperforms other methods in group detection and shared attention estimation. Additional analyses validate the effectiveness of the proposed components. Code: https://github.com/chihina/sagd-CVPRW2026.
