Research map / Reinforcement learning / Offline RL
Predicting Head Movement in Panoramic Video: A Deep Reinforcement Learning Approach
Mai Xu, Yuhang Song, Jianyi Wang, Minglang Qiao, Liangyu Huo, Zulin Wang
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 |
|---|---|---|---|
| use reinforcement learning instead of supervised regressor and classifier (ablation) | offline-DHP · PVS-HM test | CC: +0.091 average delta CC | Improved |
| use reinforcement learning instead of supervised regressor and classifier (ablation) | offline-DHP · PVS-HM test | NSS: +1.066 average delta NSS | Improved |
| use reinforcement learning instead of supervised regressor and classifier (ablation) | offline-DHP · PVS-HM test | shuffled-AUC: +0.118 average delta S-AUC | Improved |
| use reinforcement learning instead of supervised regressor and classifier (ablation) | online-DHP · PVS-HM test | MO: +0.05 average delta MO | Improved |
| initialize online-DHP with the offline DRL network (ablation) | online-DHP · PVS-HM test | MO: 0.50 -> 0.75 | Improved |
| feed ground-truth previous HM positions as input (ablation) | online-DHP · PVS-HM test | MO: 0.57 -> 0.75 | Improved |
| add the FCB (front-center bias) feature (ablation) | offline-DHP · PVS-HM test | CC: 0.654 -> 0.704 | Improved |
| add the FCB feature (ablation) | offline-DHP · PVS-HM test | NSS: 3.189 -> 3.275 | Improved |
| add the FCB feature (ablation) | SALICON · PVS-HM test | CC: 0.253 -> 0.570 | Improved |
| increase number of DRL workflows N from 1 up to 58 (ablation) | offline-DHP · PVS-HM test | CC, NSS, shuffled-AUC: converges at N >= 48 for CC and N >= 58 for NSS and shuffled-AUC | Improved |
| use DHP instead of Deep 360 Pilot (no ground-truth previous positions) | online-DHP · PVS-HM test | MO: 0.40 -> 0.57 | Improved |
| use online-DHP instead of Deep 360 Pilot | online-DHP · PVS-HM test | MO: 0.40 -> 0.75 | Improved |
Closest papers by meaning
- ViewPoint: Panoramic Video Generation with Pretrained Diffusion Models (NeurIPS 2025 · similarity 0.86)
- PanoEnv: Exploring 3D Spatial Intelligence in Panoramic Environments with Reinforcement Learning (CVPR 2026 · similarity 0.85)
- Gaze Beyond the Frame: Forecasting Egocentric 3D Visual Span (NeurIPS 2025 · similarity 0.84)
- EyEar: Learning Audio Synchronized Human Gaze Trajectory Based on Physics-Informed Dynamics (AAAI 2025 · similarity 0.84)
- PanoWorld: Geometry-Consistent Panoramic Video World Modeling (NeurIPS 2026 · similarity 0.84)
- HVIS: A Human-like Vision and Inference System for Human Motion Prediction (AAAI 2025 · similarity 0.84)
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.
