Presentation

Mapping Human Attention in Reconstructed 3D Scenes: Gaze, Language, and Point Cloud Fusion
DescriptionThis BOF explores a pipeline that fuses point clouds, natural language object descriptions, and eye gaze data to reconstruct architectural scenes and visualize human attention in 3D. Using SLAM-corrected trajectories, ray-object intersections, and interactive Plotly visualizations, the system reveals where users look and how they interact with spatial environments. Potential topics of discussion: methods for aligning multimodal data, computing gaze-object intersections, and designing interpretable heatmaps.

Contact Info: [email protected]