Research map / Detection and segmentation / Tracking
Ocean: Object-aware Anchor-free Tracking
Zhipeng Zhang, Houwen Peng, Jianlong Fu Bing Li, Weiming Hu
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 centralized sampling for classification labels (ablation) | Ocean baseline · VOT-2018 | EAO: 0.358 -> 0.396 | Improved |
| add feature combination module (multi-dilation correlation) (ablation) | Ocean baseline · VOT-2018 | EAO: 0.396 -> 0.438 | Improved |
| add object-aware classification (feature alignment) (ablation) | Ocean · VOT-2018 | EAO: 0.438 -> 0.467 | Improved |
| add plug-in online update module (ablation) | Ocean offline · VOT-2018 | EAO: 0.467 -> 0.489 | Improved |
| add a second parallel standard convolution Phi11 to feature combination (ablation) | Ocean feature combination · VOT-2018 | EAO: 0.425 -> 0.433 | Improved |
| use different dilation strides along X and Y (Phi12 instead of Phi11 as second kernel) (ablation) | Ocean feature combination · VOT-2018 | EAO: 0.433 -> 0.446 | Improved |
| add third kernel Phi21 to the Phi11 Phi12 combination (ablation) | Ocean feature combination · VOT-2018 | EAO: 0.446 -> 0.467 | Improved |
| compare Ocean with SiamRPN++ under identical training data (ablation) | Ocean vs SiamRPN++ · VOT-2018 | EAO: 0.414 -> 0.455 | Improved |
| add GOT-10k images to training (ablation) | Ocean · VOT-2018 | EAO: 0.455 -> 0.467 | Improved |
| add LaSOT to training (ablation) | Ocean · VOT-2018 | EAO: 0.467 -> 0.462 | No clear effect |
| increase number of conv layers in anchor-free networks from 0 to 3 (ablation) | Ocean anchor-free networks · OTB-100 | AUC: 0.645 -> 0.672 | Improved |
| increase conv layers from 4 to 5 (ablation) | Ocean anchor-free networks · OTB-100 | AUC: 0.673 -> 0.670 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ATOM — ATOM: Accurate Tracking by Overlap Maximization (CVPR 2019) · better on 2 of 2 shared results
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
- SiamBAN: Target-Aware Tracking With Siamese Box Adaptive Network (TPAMI 2022 · similarity 0.87)
- UAST: Uncertainty-Aware Siamese Tracking (ICML 2022 · similarity 0.85)
- Soft Anchor-Point Object Detection (ECCV 2020 · similarity 0.85)
- Bridging the Gap Between Anchor-Based and Anchor-Free Detection via Adaptive Training Sample Selection (CVPR 2020 · similarity 0.85)
- Probabilistic Anchor Assignment with IoU Prediction for Object Detection (ECCV 2020 · similarity 0.85)
- Learn To Match: Automatic Matching Network Design for Visual Tracking (ICCV 2021 · 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.
