Research map / Detection and segmentation / Tracking
D3S - A Discriminative Single Shot Segmentation Tracker
Check 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 |
|---|---|---|---|
| combine geometrically invariant segmentation cues with a geometrically constrained Euclidean target model | D3S · VOT2018 | EAO | Helped |
| remove the GIM foreground similarity channel (ablation) | D3S · VOT2018 | EAO: 0.489 -> 0.467 | Hurt |
| add the GIM target posterior channel (ablation) | D3S · VOT2018 | EAO: 0.423 -> 0.489 | Helped |
| add GIM foreground similarity and target posterior channels (ablation) | D3S · VOT2018 | EAO: 0.357 -> 0.489 | Helped |
| add the GEM target location channel (ablation) | D3S · VOT2018 | EAO: 0.251 -> 0.489 | Helped |
| update the DCF using D3S's final position estimate rather than the DCF's own estimate (ablation) | D3S · VOT2018 | EAO: 0.452 -> 0.489 | Helped |
| use the proposed rotated bounding-box fitting instead of a minimal-area rotated box (ablation) | D3S · VOT2018 | EAO: 0.444 -> 0.489 | Helped |
| use the proposed rotated bounding-box fitting instead of a min-max axis-aligned box (ablation) | D3S · VOT2018 | EAO: 0.398 -> 0.489 | Helped |
| use single-shot segmentation tracking rather than SiamMask | D3S · VOT2016 | robustness: 0.214 -> 0.131 | Helped |
| use D3S rather than SiamMask | D3S · DAVIS16 | mean Jaccard index: 71.7 -> 75.4 | Helped |
| use D3S rather than SiamRPN++ | D3S · TrackingNet test set | AUC: 73.3 -> 72.8 | Hurt |
Closest papers by meaning
- Fast Online Object Tracking and Segmentation: A Unifying Approach (CVPR 2019 · similarity 0.88)
- TrackCraft3R: Repurposing Video Diffusion Transformers for Dense 3D Tracking (NeurIPS 2026 · similarity 0.87)
- When Trackers Fail: VLM-Guided Verification and Recovery for Robust Video Object Segmentation (NeurIPS 2026 · similarity 0.86)
- GMOS: Grounding Moving Object Segmentation in 3D Space and Time (NeurIPS 2026 · similarity 0.86)
- M3SOT: Multi-Frame, Multi-Field, Multi-Space 3D Single Object Tracking (AAAI 2024 · similarity 0.85)
- Any3DIS: Class-Agnostic 3D Instance Segmentation by 2D Mask Tracking (CVPR 2025 · similarity 0.85)
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.