Research map / Detection and segmentation / Tracking
A Discriminative Single-Shot Segmentation Network for Visual Object Tracking
Alan Lukežič, Jiří Matas, Matej Kristan
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 geometrically invariant model (GIM) (ablation) | D3S2 · VOT2020 | EAO: 0.439 -> 0.513 | Improved |
| add geometrically constrained model (GEM) (ablation) | D3S2 · VOT2020 | EAO: 0.463 -> 0.513 | Improved |
| add target size estimation module (SEM) (ablation) | D3S2 · VOT2020 | EAO: 0.492 -> 0.513 | Improved |
| add SEM to preliminary D3S (ablation) | D3S · VOT2020 | EAO: 0.439 -> 0.475 | Improved |
| add target size estimation module (SEM) (ablation) | D3S2 · LaSoT | AUC: 49.2 -> 54.0 | Improved |
| add mask adjustment module instead of bilinear resampling (ablation) | D3S2 · VOT2020 | EAO: 0.496 -> 0.513 | Improved |
| add segmentation mask to SEM (ablation) | D3S2 · VOT2020 | EAO: 0.508 -> 0.513 | Improved |
| add channel attention in refinement pathway (ablation) | D3S2 · VOT2020 | EAO: 0.497 -> 0.513 | Improved |
| use trainable MLP in GIM instead of average top-K (ablation) | D3S2 · VOT2020 | EAO: 0.468 -> 0.513 | Improved |
| use 256 DCF channels instead of 64 (ablation) | D3S2 · VOT2020 | EAO: 0.504 -> 0.513 | Improved |
| use octave-convolution ResNet50 instead of vanilla ResNet50 (ablation) | D3S2 · VOT2020 | EAO: 0.479 -> 0.513 | Improved |
| D3S2 versus preliminary D3S | D3S2 · DAVIS16 | mean Jaccard index: 75.4 -> 74.2 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- Siam R-CNN — Siam R-CNN: Visual Tracking by Re-Detection (CVPR 2020) · better on 1 of 5 shared results
- ATOM — ATOM: Accurate Tracking by Overlap Maximization (CVPR 2019) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- D3S - A Discriminative Single Shot Segmentation Tracker (CVPR 2020 · similarity 0.98)
- Fast Online Object Tracking and Segmentation: A Unifying Approach (CVPR 2019 · similarity 0.90)
- Robust Visual Tracking by Segmentation (ECCV 2022 · similarity 0.89)
- Visual Object Tracking with Discriminative Filters and Siamese Networks: A Survey and Outlook (TPAMI 2022 · similarity 0.89)
- When Trackers Fail: VLM-Guided Verification and Recovery for Robust Video Object Segmentation (NeurIPS 2026 · similarity 0.87)
- TrackCraft3R: Repurposing Video Diffusion Transformers for Dense 3D Tracking (NeurIPS 2026 · 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.
