Research map / Video / Video segmentation and tracking
Tracking through Severe Occlusion via Event-Derived Transient Cues
Hao Dong, Yujin Liu, Haoyue Liu, Zhenyu Wang, Shihan Peng, Zhiwei Shi, Yi Chang, Luxin Yan
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 bidirectional time-surface with motion-consistency supervision to forward-only motion autoregression (ablation) | EvoTrack · Ablation (FE108-matching) | PR / SR: 92.1 -> 94.6 PR; 66.9 -> 68.4 SR | Improved |
| combine appearance matching (TAM) with motion autoregression (EMA) (ablation) | EvoTrack · Ablation | PR / SR | Improved |
| use Gaussian masking instead of random masking (ablation) | TAM · Masking ablation | SR: 60.2 -> 62.1 | Improved |
| use random masking instead of no masking (ablation) | TAM · Masking ablation | PR: 75.7 -> 79.7 | Improved |
| use Gaussian masking instead of random masking (ablation) | TAM · Masking ablation | PR: 79.7 -> 80.1 | Improved |
| use gated adaptive fusion instead of addition (ablation) | EvoTrack · Fusion ablation | PR: 77.5 -> 80.1 | Improved |
| concatenate instead of add motion and appearance features (ablation) | EvoTrack · Fusion ablation | PR: 77.5 -> 78.9 | Improved |
| use gated fusion instead of concatenation (ablation) | EvoTrack · Fusion ablation | SR: 62.0 -> 62.1 | No clear effect |
| use EvoTrack vs SeqTrack v2 on VisEvent | EvoTrack · VisEvent | SR: 63.0 (SeqTrack v2) vs 62.1 | Hurt |
| add event-based motion cues for occlusion tracking | EvoTrack · FEOT (occlusion) | PR: 58.0 (SDSTrack) -> 62.7 | Improved |
| increase occlusion ratio above 60% | EvoTrack · FEOT | SR | Hurt |
Closest papers by meaning
- Motion-Aware Object Tracking via Motion and Geometry-Aware Cues (AAAI 2026 · similarity 0.88)
- TimeTracker: Event-based Continuous Point Tracking for Video Frame Interpolation with Non-linear Motion (CVPR 2025 · similarity 0.86)
- Autoregressive Sequential Pretraining for Visual Tracking (CVPR 2025 · similarity 0.86)
- ETAP: Event-based Tracking of Any Point (CVPR 2025 · similarity 0.85)
- Tracktention: Leveraging Point Tracking to Attend Videos Faster and Better (CVPR 2025 · similarity 0.85)
- Transient Glimpses: Unveiling Occluded Backgrounds through the Spike Camera (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.
