Research map / Detection and segmentation / Tracking
End-to-end Active Object Tracking via Reinforcement Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 randomized environment augmentation during training (ablation) | Active tracker · Standard | AR: 840$\pm$602 -> 2547$\pm$58 | Helped |
| Add randomized environment augmentation during training (ablation) | Active tracker · S2MP2 | EL: 461.9$\pm$180.0 -> 2603.6$\pm$503.0 | Helped |
| Use end-to-end active tracking instead of passive tracking plus PID-like camera control | ViZDoom trackers · Standard, SharpTurn, and Cacodemon | AR: -454.2 $\pm$ 0.3 -> 2457$\pm$58; -454.4 $\pm$ 0.3 -> 2560$\pm$34; -453.5 $\pm$ 0.2 -> 2451$\pm$71 | Helped |
| Use end-to-end active tracking instead of passive tracking plus camera control | UE trackers · S1SP1 | AR: -453.8$ 5cpm$0.8 -> 2495.7 | Helped |
| Use end-to-end active tracking instead of passive tracking plus camera control | UE trackers · S1MP1 | AR: 708.3$\pm$3.3 -> 2106.0 | Helped |
| Use end-to-end active tracking instead of passive tracking plus camera control | UE trackers · S1SP2 | AR: -452.4$\pm$0.7 -> 2162.5 | Helped |
| Change the target appearance, background, movement path, and add distractors | RandomizedEnv active tracker · ViZDoom unseen environments | AR: 2415$\pm$71; 2386$\pm$86; 1504 $\pm$ 158; 2636 $\pm$ 34; 2560$\pm$34; 2537$\pm$58; 2493$\pm$72; 2313$\pm$103 | Mixed |
| Use an LSTM to incorporate historical observations | Active tracker · SharpTurn | tracking recovery | Helped |
| Train in a simulator and evaluate predicted camera actions on recorded real-world video | UE-trained active tracker · VOT Woman and Sphere clips | action consistency | Helped |
Closest papers by meaning
- Visual Tracking via Hierarchical Deep Reinforcement Learning (AAAI 2021 · similarity 0.88)
- Online Decision Based Visual Tracking via Reinforcement Learning (NeurIPS 2020 · similarity 0.88)
- Action-Decision Networks for Visual Tracking With Deep Reinforcement Learning (CVPR 2017 · similarity 0.88)
- Open-World Drone Active Tracking with Goal-Centered Rewards (NeurIPS 2025 · similarity 0.86)
- Track-On: Transformer-based Online Point Tracking with Memory (ICLR 2025 · similarity 0.84)
- Video Diffusion Models Excel at Tracking Similar-Looking Objects Without Supervision (NeurIPS 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.