Research map / Video / Video segmentation and tracking
Deep Future Gaze: Gaze Anticipation on Egocentric Videos Using Adversarial Networks
Mengmi Zhang, Keng Teck Ma, Joo Hwee Lim, Qi Zhao, Jiashi Feng
Check 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 |
|---|---|---|---|
| Use DFG for gaze anticipation | GTEA · GTEA | AUC: 0.878 -> 0.883 | Improved |
| Use DFG for gaze anticipation | GTEA · GTEA | AAE: 8.4 -> 10.5 | Hurt |
| Use DFG for gaze anticipation | GTEAplus · GTEAplus | AUC: 0.867 -> 0.952 | Improved |
| Use DFG for gaze anticipation | GTEAplus · GTEAplus | AAE: 7.9 -> 6.6 | Improved |
| Use DFG for gaze anticipation | OST · OST | AUC: 0.850 -> 0.854 | Improved |
| Use DFG for gaze anticipation | OST · OST | AAE: 13.3 -> 10.6 | Improved |
| Integrate anticipated gaze locations into C3D | C3D on GTEAplus · GTEAplus | activity recognition rate: 26.9% -> 28.5% | Improved |
| Add a moving-background stream via the two-stream 3D-CNN generator (ablation) | DFG on GTEA · GTEA | AAE: 12.3, 11.9, 12.2, 12.1; Our Best: 11.0, 11.3, 11.3, 11.5 | Improved |
| Train GP end-to-end on generated frames rather than only real frames (ablation) | DFG on GTEA · GTEA | AAE: 12.8, 13.3, 13.9, 13.9; Our Best: 11.0, 11.3, 11.3, 11.5 | Improved |
| Add the discriminator to provide realism feedback during frame generation (ablation) | DFG on GTEA · GTEA | AAE: 11.3, 11.8, 12.3, 12.4; Our Best: 11.0, 11.3, 11.3, 11.5 | Improved |
| Add the two-stream 3D-CNN generator instead of the static-background variant (ablation) | DFG on GTEA · GTEA | AUC: 0.86, 0.85, 0.85, 0.85; Our Best: 0.88, 0.86, 0.85, 0.84 | Mixed |
| Train GP end-to-end on generated frames rather than only real frames (ablation) | DFG on GTEA · GTEA | AUC: 0.84, 0.82, 0.80, 0.81; Our Best: 0.88, 0.86, 0.85, 0.84 | Improved |
Closest papers by meaning
- Anticipating Where People will Look Using Adversarial Networks (TPAMI 2018 · similarity 0.97)
- Generative Adversarial Network for Future Hand Segmentation from Egocentric Video (ECCV 2022 · similarity 0.90)
- Gaze Beyond the Frame: Forecasting Egocentric 3D Visual Span (NeurIPS 2025 · similarity 0.89)
- GazeFlow: From Human Gaze Behavior to Generative Egocentric Gaze Prediction (NeurIPS 2026 · similarity 0.88)
- Gaze-VLM: Bridging Gaze and VLMs through Attention Regularization for Egocentric Understanding (NeurIPS 2025 · similarity 0.86)
- Exocentric-to-Egocentric Video Generation (NeurIPS 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.
