Research map / Video / Video segmentation and tracking
RANet: Ranking Attention Network for Fast Video Object Segmentation
Ziqin Wang, Jun Xu, Li Liu, Fan Zhu, 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 |
|---|---|---|---|
| add the ranking attention module (versus keeping all similarity maps unranked) (ablation) | RANet · DAVIS16 val | J Mean: 81.9 -> 85.5 | Improved |
| add ranking attention instead of channel-wise maximum over similarity maps (ablation) | RANet · DAVIS16 val | J Mean: 81.1 -> 85.5 | Improved |
| add the correlation layer (ablation) | RANet · DAVIS16 val | J Mean: 67.5 -> 85.5 | Improved |
| add the previous frame's mask as input (ablation) | RANet · DAVIS16 val | J Mean: 81.4 -> 85.5 | Improved |
| add static image pre-training (ablation) | RANet · DAVIS16 val | J Mean: 73.2 -> 85.5 | Improved |
| add video fine-tuning (ablation) | RANet · DAVIS16 val | J Mean: 79.9 -> 85.5 | Improved |
| add online learning iterations at test time | RANet · DAVIS16 val | J&F Mean: 85.5 -> 87.1 | Improved |
| add online learning at test time (runtime cost) | RANet · DAVIS16 val | time per frame (s): 0.033 -> 4.00 | Hurt |
| add online learning when trained only on static images | RANet · DAVIS16 trainval (static images only, no video fine-tuning) | J Mean: 83.2 -> 86.2 | Improved |
| replace RGMP-style feature concatenation with correlation plus ranking attention | RGMP · DAVIS16 val | J Mean: 81.5 -> 85.5 | Improved |
Later papers that compared with this method
5 later accepted papers report RaNet as a baseline in their results tables; 5 of them report a better number on most of the results they share with it.
- MomentDiff: Generative Video Moment Retrieval from Random to Real (NeurIPS 2023 · better on 1 of 1 shared result)
- Spatiotemporal Graph Neural Network based Mask Reconstruction for Video Object Segmentation (AAAI 2021 · better on 1 of 1 shared result)
- State-Aware Tracker for Real-Time Video Object Segmentation (CVPR 2020 · better on 1 of 1 shared result)
- Fast Video Object Segmentation using the Global Context Module (ECCV 2020 · better on 1 of 1 shared result)
- Kernelized Memory Network for Video Object Segmentation (ECCV 2020 · better on 1 of 1 shared result)
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.
- FEELVOS — FEELVOS: Fast End-To-End Embedding Learning for Video Object Segmentation (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
- Learning Unsupervised Video Object Segmentation Through Visual Attention (CVPR 2019 · similarity 0.87)
- Find Objects and Focus on Highlights: Mining Object Semantics for Video Highlight Detection via Graph Neural Networks (AAAI 2020 · similarity 0.85)
- Patchwork: A Patch-Wise Attention Network for Efficient Object Detection and Segmentation in Video Streams (ICCV 2019 · similarity 0.85)
- Guided Slot Attention for Unsupervised Video Object Segmentation (CVPR 2024 · similarity 0.84)
- Putting the Object Back into Video Object Segmentation (CVPR 2024 · similarity 0.83)
- Dual Prototype Attention for Unsupervised Video Object Segmentation (CVPR 2024 · similarity 0.82)
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.
