Papers › SeqFormer: Sequential Transformer for Video Instance Segmentation

SeqFormer: Sequential Transformer for Video Instance Segmentation

15 Dec 2021arXiv:2112.08275archive 2025-07-28

Junfeng Wu, Yi Jiang, Song Bai, Wenqing Zhang, Xiang Bai

In this work, we present SeqFormer for video instance segmentation. SeqFormer follows the principle of vision transformer that models instance relationships among video frames. Nevertheless, we observe that a stand-alone instance query suffices for capturing a time sequence of instances in a video, but attention mechanisms shall be done with each frame independently. To achieve this, SeqFormer locates an instance in each frame and aggregates temporal information to learn a powerful representation of a video-level instance, which is used to predict the mask sequences on each frame dynamically. Instance tracking is achieved naturally without tracking branches or post-processing. On YouTube-VIS, SeqFormer achieves 47.4 AP with a ResNet-50 backbone and 49.0 AP with a ResNet-101 backbone without bells and whistles. Such achievement significantly exceeds the previous state-of-the-art performance by 4.6 and 4.4, respectively. In addition, integrated with the recently-proposed Swin transformer, SeqFormer achieves a much higher AP of 59.3. We hope SeqFormer could be a strong baseline that fosters future research in video instance segmentation, and in the meantime, advances this field with a more robust, accurate, neat model. The code is available at https://github.com/wjf5203/SeqFormer.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

wjf5203/SeqFormer officialmentioned in papermentioned on GitHubpytorchNOASSERTION report
wjf5203/vnext mentioned on GitHubpytorchApache-2.0 report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Instance SegmentationSemantic SegmentationVideo Instance Segmentation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Video Instance Segmentation HQ-YTVIS SeqFormer (Swin-L) Tube-Boundary AP 43.3 #2 of 4 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (Swin-L) AP50 82.1 #9 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (Swin-L) AP75 66.4 #9 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (Swin-L) AR1 51.7 #9 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (Swin-L) AR10 64.4 #9 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (Swin-L) mask AP 59.3 #9 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (ResNet-101) AP50 71.1 #17 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (ResNet-101) AP75 55.7 #17 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (ResNet-101) AR1 46.8 #17 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (ResNet-101) AR10 56.9 #17 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (ResNet-101) mask AP 49.0 #17 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (ResNet-50) AP50 69.8 #19 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (ResNet-50) AP75 51.8 #19 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (ResNet-50) AR1 45.5 #19 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (ResNet-50) AR10 54.8 #19 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (ResNet-50) mask AP 47.4 #19 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (ResNet-50) AP50 66.9 #22 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (ResNet-50) AP75 50.5 #22 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (ResNet-50) AR1 45.6 #22 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (ResNet-50) AR10 54.6 #22 of 44 Archive leaderboard report
Video Instance Segmentation YouTube-VIS validation SeqFormer (ResNet-50) mask AP 45.1 #22 of 44 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

AttentionDense ConnectionsLayer NormalizationLinear LayerMulti-Head AttentionResidual ConnectionSoftmaxVision Transformer

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections