Research map / Detection and segmentation / Tracking
Fast Online Object Tracking and Segmentation: A Unifying Approach
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 |
|---|---|---|---|
| Add a segmentation mask branch and train it jointly with similarity and box prediction (ablation) | SiamRPN · VOT-2018 | EAO: 0.359 -> 0.380 | Helped |
| Add a segmentation mask branch to SiamFC, without box regression (ablation) | SiamFC · VOT-2018 | EAO: 0.251 to 0.265 | Helped |
| Use the mask branch rather than the box branch to generate boxes (ablation) | SiamMask · VOT-2018 | EAO: 0.363 -> 0.380 | Helped |
| Use a rotated minimum bounding rectangle rather than an axis-aligned Min-max box (ablation) | SiamMask · VOT-2016 random crops | mAP@0.7 IOU | Helped |
| Use Opt box generation instead of MBR (ablation) | SiamMask · VOT-2018 | EAO: 0.380 -> 0.387 | Helped |
| Use Opt box generation instead of MBR (ablation) | SiamMask · VOT-2018 | Speed: 55 to 5 | Hurt |
| Remove the box regression branch (ablation) | SiamMask · VOT-2018 | EAO: 0.380 -> 0.334 | Hurt |
| Remove the mask refinement modules (ablation) | SiamMask · DAVIS-2016 validation | F mean: 67.8 -> 57.8 | Hurt |
| Replace AlexNet with ResNet-50 as the shared backbone (ablation) | SiamRPN · VOT-2018 | EAO: 0.243 -> 0.359 | Helped |
| Use MBR mask-derived boxes rather than axis-aligned box representation (ablation) | SiamMask · VOT-2016 random crops | mAP@0.5 IOU: +29 and +9.2 points | Helped |
Closest papers by meaning
- Fast Video Object Segmentation by Reference-Guided Mask Propagation (CVPR 2018 · similarity 0.91)
- Fast and Accurate Online Video Object Segmentation via Tracking Parts (CVPR 2018 · similarity 0.89)
- Siamese Network with Interactive Transformer for Video Object Segmentation (AAAI 2022 · similarity 0.89)
- SiamCAR: Siamese Fully Convolutional Classification and Regression for Visual Tracking (CVPR 2020 · similarity 0.89)
- Siam R-CNN: Visual Tracking by Re-Detection (CVPR 2020 · similarity 0.89)
- High Performance Visual Tracking With Siamese Region Proposal Network (CVPR 2018 · similarity 0.89)
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.