Research map / Video / Video segmentation and tracking
Full-Duplex Strategy for Video Object Segmentation
Ge-Peng Ji, Keren Fu, Zhe Wu, Deng-Ping Fan, Jianbing Shen, Ling Shao
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 |
|---|---|---|---|
| fuse appearance and motion with RCAM instead of vanilla concat-conv (ablation) | FSNet · DAVIS16 | S-alpha: 0.871 -> 0.900 | Improved |
| fuse appearance and motion with RCAM instead of vanilla concat-conv (ablation) | FSNet · MCL | S-alpha: 0.776 -> 0.833 | Improved |
| add BPM on top of RCAM (ablation) | FSNet · DAVIS16 | S-alpha: 0.900 -> 0.920 | Improved |
| add BPM on top of RCAM (ablation) | FSNet · MCL | S-alpha: 0.833 -> 0.864 | Improved |
| add BPM to vanilla fusion (Bi-Purf.) (ablation) | vanilla fusion · MCL | S-alpha: 0.776 -> 0.855 | Improved |
| add RCAM to BPM-only model (ablation) | FSNet · DAVIS16 | S-alpha: 0.904 -> 0.920 | Improved |
| use motion maps instead of appearance frames as single stimulus (ablation) | single-branch model · DAVIS16 | S-alpha: 0.834 -> 0.858 | Improved |
| use motion instead of appearance as single stimulus (ablation) | single-branch model · MCL | MAE: 0.038 -> 0.053 | Mixed |
| use full-duplex RCAM and BPM instead of best simplex variant (ablation) | FSNet · DAVIS16 | S-alpha: 0.902 -> 0.920 | Improved |
| use bidirectional BPM interaction instead of self-purification (ablation) | FSNet · DAVIS16 | S-alpha: 0.899 -> 0.920 | Improved |
| increase number of BPMs from 0 to 4 (ablation) | FSNet · DAVIS16 | S-alpha: 0.900 -> 0.920 | Improved |
| increase BPMs from 4 to 6 (ablation) | FSNet · DAVIS16 | S-alpha: 0.920 -> 0.918 | No clear effect |
Closest papers by meaning
- VOSTR: Video Object Segmentation via Transferable Representations (IJCV 2020 · similarity 0.86)
- Unsupervised Object Segmentation in Video by Efficient Selection of Highly Probable Positive Features (ICCV 2017 · similarity 0.85)
- Depth-aware Test-Time Training for Zero-shot Video Object Segmentation (CVPR 2024 · similarity 0.84)
- A Survey on Deep Learning Technique for Video Segmentation (TPAMI 2022 · similarity 0.84)
- A Benchmark Dataset and Evaluation Methodology for Video Object Segmentation (CVPR 2016 · similarity 0.84)
- SAM2TAP: Repurposing Pre-trained Video Object Segmentation Models for Point Tracking (IJCV 2026 · similarity 0.83)
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.
