Research map / Detection and segmentation / CNN features
Object-aware Gaze Target Detection
Francesco Tonini, Nicola Dall'Asen, Cigdem Beyan, Elisa Ricci
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add Gaze Object Transformer on top of detector features (ablation) | Object-aware gaze target detector · GazeFollow | AUC: 0.864 -> 0.918 | Improved |
| add object masking (queries from heads only) (ablation) | GOT · GazeFollow | Avg. distance: 0.075 -> 0.073 | Improved |
| add gaze cone without no cone-object skip (ablation) | GOT with object masking · GazeFollow | AUC: 0.919 -> 0.905 | Hurt |
| add no cone-object skip together with gaze cone (ablation) | full method · GazeFollow | AUC: 0.905 -> 0.922 | Improved |
| add no cone-object skip together with gaze cone (ablation) | full method · VideoAttentionTarget | Dist.: 0.154 -> 0.101 | Improved |
| use 3D gaze cone instead of 2D (ablation) | Object-aware gaze target detector · GazeFollow | Avg. angular error: 7.7 -> 7.2 | Improved |
| 3D versus 2D cone under high annotation variance (ablation) | Object-aware gaze target detector · GazeFollow AUC at 10% variance retained | AUC: 0.980 vs 0.980 | No clear effect |
| add separate DETR detector to our model for gazed-object detection (ablation) | Object-aware gaze target detector · GazeFollow COCO subset | AP: 0.14 -> 0.03 | Hurt |
| object-aware transformer versus Tu et al. under real head detections | Object-aware gaze target detector · VideoAttentionTarget | Out-of-frame AP: 0.854 -> 0.944 | Improved |
Closest papers by meaning
- TransGOP: Transformer-Based Gaze Object Prediction (AAAI 2024 · similarity 0.90)
- Gaze-LLE: Gaze Target Estimation via Large-Scale Learned Encoders (CVPR 2025 · similarity 0.89)
- Gaze Target Detection by Merging Human Attention and Activity Cues (AAAI 2024 · similarity 0.88)
- Toward Semantic Gaze Target Detection (NeurIPS 2024 · similarity 0.87)
- Unifying Top-down and Bottom-up Scanpath Prediction Using Transformers (CVPR 2024 · similarity 0.85)
- Sharingan: A Transformer Architecture for Multi-Person Gaze Following (CVPR 2024 · similarity 0.85)
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.
