Research map / Detection and segmentation / Tracking
Towards Universal Modal Tracking With Online Dense Temporal Token Learning
Yaozong Zheng, Bineng Zhong, Qihua Liang, Shengping Zhang, Guorong Li, Xianxian Li, Rongrong Ji
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 temporal token association (ablation) | ODTrack · LaSOT | AUC: 71.0 -> 72.8 | Improved |
| use concatenation instead of separate token association (ablation) | ODTrack · LaSOT | AUC: 72.2 -> 72.8 | Improved |
| increase search clip length from 2 to 3 (ablation) | ODTrack · LaSOT | AUC: 72.8 -> 73.1 | Improved |
| increase search clip length from 3 to 5 (ablation) | ODTrack · LaSOT | AUC: 73.1 -> 72.0 | Hurt |
| increase sampling range from 200 to 400 (ablation) | ODTrack · LaSOT | AUC: 72.8 -> 73.2 | Improved |
| add conditional gate (ablation) | UM-ODTrack · DepthTrack | F-score (%): 66.7 -> 68.0 | Improved |
| add gated modal-scalable perceiver (GMP) (ablation) | UM-ODTrack · DepthTrack | F-score (%): 68.0 -> 69.3 | Improved |
| use tanh instead of ReLU gate activation (ablation) | UM-ODTrack · DepthTrack | F-score (%): 65.0 -> 69.3 | Improved |
| use tanh instead of sigmoid gate activation (ablation) | UM-ODTrack · DepthTrack | F-score (%): 67.0 -> 69.3 | Improved |
| add depth/thermal/event modality to RGB (ablation) | UM-ODTrack · LasHeR | SR: 54.5 -> 60.8 | Improved |
| share modality tokenizers (ablation) | UM-ODTrack · LasHeR | SR: 48.5 -> 60.8 | Improved |
| add one-shot multi-task joint training (ablation) | UM-ODTrack · DepthTrack | F-score (%): 67.8 -> 69.3 | Improved |
Closest papers by meaning
- Single-Model and Any-Modality for Video Object Tracking (CVPR 2024 · similarity 0.90)
- Exploring Modality-Aware Fusion and Decoupled Temporal Propagation for Multi-Modal Object Tracking (AAAI 2026 · similarity 0.90)
- ODTrack: Online Dense Temporal Token Learning for Visual Tracking (AAAI 2024 · similarity 0.90)
- Collaborating Vision, Depth, and Thermal Signals for Multi-Modal Tracking: Dataset and Algorithm (NeurIPS 2025 · similarity 0.89)
- Cross-Modal Object Tracking: Modality-Aware Representations and a Unified Benchmark (AAAI 2022 · similarity 0.89)
- Local-Global Multi-Modal Distillation for Weakly-Supervised Temporal Video Grounding (AAAI 2024 · similarity 0.87)
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.
