Research map / Video / Video segmentation and tracking
A Benchmark Dataset and Evaluation Methodology for Video Object Segmentation
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 |
|---|---|---|---|
| Use NLC rather than the other evaluated algorithms | Video object segmentation on DAVIS · DAVIS | J mean: 0.641 | Helped |
| Use FCP rather than sequential propagation methods | Video object segmentation on DAVIS · DAVIS | J decay: 0.031 | Helped |
| Propagate segmentation across consecutive frames | JMP and SEA · DAVIS | T mean: JMP: 0.131; SEA: 0.137 | Helped |
| Use motion cues rather than color cues for saliency preprocessing | Saliency preprocessing on DAVIS · DAVIS | J mean: 0.173 -> 0.532 | Helped |
| Use large-range connectivity | NLC, FCP, and KEY · DAVIS, occlusion | J: NLC: 0.70 -0.09; FCP: 0.59 +0.07 | Helped |
| Rely on distinctive motion saliency | NLC and SAL · DAVIS, dynamic background | J: NLC: 0.53 +0.15; SAL: 0.35 +0.10 | Hurt |
| Use closed motion-boundary curves that follow object contours | FST · DAVIS, dynamic background | J: 0.53 +0.06 | Helped |
| Use optical-flow-dependent motion information under fast-motion conditions | Methods exploiting motion information · DAVIS, fast motion | J | Hurt |
| Use non-local connectivity in the random-walk matrix | NLC · DAVIS, fast motion | J: 0.64 +0.00 | Helped |
| Propagate annotations sequentially between frames | SEA and JMP · DAVIS, occlusion | J: SEA: 0.47 +0.13; JMP: 0.47 +0.21 | Hurt |
Closest papers by meaning
- From ViT Features to Training-free Video Object Segmentation via Streaming-data Mixture Models (NeurIPS 2023 · similarity 0.88)
- Breaking the "Object" in Video Object Segmentation (CVPR 2023 · similarity 0.87)
- Video K-Net: A Simple, Strong, and Unified Baseline for Video Segmentation (CVPR 2022 · similarity 0.87)
- HODOR: High-Level Object Descriptors for Object Re-Segmentation in Video Learned From Static Images (CVPR 2022 · similarity 0.87)
- Learning Spatial-Semantic Features for Robust Video Object Segmentation (ICLR 2025 · similarity 0.86)
- Multi-Granularity Video Object Segmentation (AAAI 2025 · similarity 0.86)
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.