Research map / Video / Video segmentation and tracking
SeqFormer: Sequential Transformer for Video Instance Segmentation
Junfeng Wu, Yi Jiang, Song Bai, Wenqing Zhang, Xiang Bai
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 instance query decomposition into frame-level box queries (ablation) | SeqFormer · YouTube-VIS 2019 val | AP: 34.1 -> 45.1 | Improved |
| keep spatial and temporal dimensions separate instead of flattening them (ablation) | SeqFormer · YouTube-VIS 2019 val | AP: 35.1 -> 42.5 | Improved |
| add multi-scale feature maps (ablation) | SeqFormer (single-scale) · YouTube-VIS 2019 val | AP: 42.5 -> 45.1 | Improved |
| use weighted-sum temporal aggregation instead of sum (ablation) | SeqFormer · YouTube-VIS 2019 val | AP: 30.6 -> 45.1 | Improved |
| use weighted-sum temporal aggregation instead of average (ablation) | SeqFormer · YouTube-VIS 2019 val | AP: 43.2 -> 45.1 | Improved |
| generate mask head from all frames instead of 1 frame (ablation) | SeqFormer · YouTube-VIS 2019 val | AP: 38.1 -> 45.1 | Improved |
| use 5 frames instead of all frames for the instance representation (ablation) | SeqFormer · YouTube-VIS 2019 val | AP: 44.6 -> 45.1 | No clear effect |
| train with COCO pseudo videos (joint training) (ablation) | SeqFormer · YouTube-VIS 2019 val | AP: 45.1 -> 47.4 | Improved |
| evaluate clip-wise with clip matching instead of whole video | SeqFormer · YouTube-VIS 2019 val | AP: 45.1 -> 42.0 | Hurt |
| use a stronger Swin-L backbone instead of ResNet-50 | SeqFormer · YouTube-VIS 2019 val | AP: 47.4 -> 59.3 | Improved |
| use SeqFormer instead of IFC | SeqFormer · YouTube-VIS 2021 val | AP: 36.6 -> 40.5 | Improved |
Later papers that compared with this method
6 later accepted papers report SeqFormer as a baseline in their results tables; 5 of them report a better number on most of the results they share with it.
- Video Instance Segmentation in an Open-World (IJCV 2024 · better on 1 of 1 shared result)
- InstanceFormer: An Online Video Instance Segmentation Framework (AAAI 2023 · better on 0 of 1 shared result)
- Video Instance Segmentation via Multi-Scale Spatio-Temporal Split Attention Transformer (ECCV 2022 · better on 16 of 16 shared results)
- Video Mask Transfiner for High-Quality Video Instance Segmentation (ECCV 2022 · better on 12 of 12 shared results)
- In Defense of Online Models for Video Instance Segmentation (ECCV 2022 · better on 3 of 3 shared results)
- MinVIS: A Minimal Video Instance Segmentation Framework without Video-based Training (NeurIPS 2022 · 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 2 of 2 on most shared results.
- CompFeat — CompFeat: Comprehensive Feature Aggregation for Video Instance Segmentation (AAAI 2021) · better on 1 of 1 shared result
- STEm-Seg — STEm-Seg: Spatio-temporal Embeddings for Instance Segmentation in Videos (ECCV 2020) · 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
- InstanceFormer: An Online Video Instance Segmentation Framework (AAAI 2023 · similarity 0.91)
- End-to-End Video Instance Segmentation With Transformers (CVPR 2021 · similarity 0.89)
- Temporally Efficient Vision Transformer for Video Instance Segmentation (CVPR 2022 · similarity 0.89)
- VideoCutLER: Surprisingly Simple Unsupervised Video Instance Segmentation (CVPR 2024 · similarity 0.86)
- Efficient Temporal Action Segmentation via Boundary-aware Query Voting (NeurIPS 2024 · similarity 0.86)
- MobileInst: Video Instance Segmentation on the Mobile (AAAI 2024 · 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.
