Research map / Video / Video segmentation and tracking
Scanpath Prediction in Panoramic Videos via Expected Code Length Minimization
Mu Li, Kanglong Fan, Kede Ma
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 historical path context to visual-only model (ablation) | Ours · CVPR18 | maxTC: 0.517 -> 0.623 | Improved |
| add causal path context (ablation) | Ours (Visual + H-Path) · CVPR18 | maxTC: 0.623 -> 0.708 | Improved |
| add causal path context (ablation) | Ours (Visual + H-Path) · VRW23 | maxTC: 0.712 -> 0.796 | Improved |
| add visual context to path-only model (ablation) | Ours · VRW23 | maxTC: 0.687 -> 0.712 | Improved |
| use expected code length instead of MSE (relative uv) (ablation) | Ours · CVPR18 | maxTC: 0.449 -> 0.708 | Improved |
| use expected code length instead of MAE (relative uv) (ablation) | Ours · VRW23 | maxTC: 0.707 -> 0.796 | Improved |
| use relative uv instead of spherical coordinates (expected code length loss) (ablation) | Ours · CVPR18 | maxTC: 0.656 -> 0.708 | Improved |
| use relative uv instead of 3D Euclidean (expected code length loss) (ablation) | Ours · VRW23 | maxTC: 0.748 -> 0.796 | Improved |
| use PID controller sampler instead of beam search (Visual + H-Path + C-Path) (ablation) | Ours · one-second videos | maxTC: 0.349 -> 0.844 | Improved |
| use PID controller sampler instead of max sampler (Visual + H-Path + C-Path) (ablation) | Ours · one-second videos | maxTC: 0.360 -> 0.844 | Improved |
| add causal path context under max sampler (ablation) | Ours (Visual + H-Path) · one-second videos | maxTC: 0.451 -> 0.360 | Hurt |
| add causal path context under PID sampler (ablation) | Ours (Visual + H-Path) · one-second videos | maxTC: 0.786 -> 0.844 | Improved |
Closest papers by meaning
- Learned Scanpaths Aid Blind Panoramic Video Quality Assessment (CVPR 2024 · similarity 0.88)
- PanoWorld: Geometry-Consistent Panoramic Video World Modeling (NeurIPS 2026 · similarity 0.86)
- ViewPoint: Panoramic Video Generation with Pretrained Diffusion Models (NeurIPS 2025 · similarity 0.85)
- Splatter-360: Generalizable 360 Gaussian Splatting for Wide-baseline Panoramic Images (CVPR 2025 · similarity 0.85)
- PanoWan: Lifting Diffusion Video Generation Models to 360$^\circ$ with Latitude/Longitude-aware Mechanisms (NeurIPS 2025 · similarity 0.84)
- SpFormer: Spatio-Temporal Modeling for Scanpaths with Transformer (AAAI 2024 · 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.
