Papers › SeMask: Semantically Masked Transformers for Semantic Segmentation

SeMask: Semantically Masked Transformers for Semantic Segmentation

23 Dec 2021arXiv 2021 12arXiv:2112.12782archive 2025-07-28

Jitesh Jain, Anukriti Singh, Nikita Orlov, Zilong Huang, Jiachen Li, Steven Walton, Humphrey Shi

Finetuning a pretrained backbone in the encoder part of an image transformer network has been the traditional approach for the semantic segmentation task. However, such an approach leaves out the semantic context that an image provides during the encoding stage. This paper argues that incorporating semantic information of the image into pretrained hierarchical transformer-based backbones while finetuning improves the performance considerably. To achieve this, we propose SeMask, a simple and effective framework that incorporates semantic information into the encoder with the help of a semantic attention operation. In addition, we use a lightweight semantic decoder during training to provide supervision to the intermediate semantic prior maps at every stage. Our experiments demonstrate that incorporating semantic priors enhances the performance of the established hierarchical encoders with a slight increase in the number of FLOPs. We provide empirical proof by integrating SeMask into Swin Transformer and Mix Transformer backbones as our encoder paired with different decoders. Our framework achieves a new state-of-the-art of 58.25% mIoU on the ADE20K dataset and improvements of over 3% in the mIoU metric on the Cityscapes dataset. The code and checkpoints are publicly available at https://github.com/Picsart-AI-Research/SeMask-Segmentation .

PaperPDFConference PDFCode

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

Code

Picsart-AI-Research/SeMask-Segmentation officialmentioned in papermentioned on GitHubpytorchNOASSERTION 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

DecoderSemantic Segmentation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Semantic Segmentation ADE20K SeMask (SeMask Swin-L FaPN-Mask2Former) Validation mIoU 58.2 #23 of 235 Archive leaderboard report
Semantic Segmentation ADE20K SeMask (SeMask Swin-L MSFaPN-Mask2Former) Validation mIoU 58.2 #24 of 235 Archive leaderboard report
Semantic Segmentation ADE20K SeMask (SeMask Swin-L Mask2Former) Validation mIoU 57.5 #32 of 235 Archive leaderboard report
Semantic Segmentation ADE20K SeMask(SeMask Swin-L MSFaPN-Mask2Former, single-scale) Validation mIoU 57.0 #36 of 235 Archive leaderboard report
Semantic Segmentation ADE20K SeMask (SeMask Swin-L MaskFormer) Validation mIoU 56.2 #41 of 235 Archive leaderboard report
Semantic Segmentation ADE20K SeMask (SeMask Swin-L FPN) Validation mIoU 53.52 #79 of 235 Archive leaderboard report
Semantic Segmentation ADE20K SeMask (SeMask Swin-B FPN) Params (M) 96 #106 of 235 Archive leaderboard report
Semantic Segmentation ADE20K SeMask (SeMask Swin-B FPN) Validation mIoU 50.98 #106 of 235 Archive leaderboard report
Semantic Segmentation ADE20K SeMask (SeMask Swin-S FPN) Params (M) 56 #159 of 235 Archive leaderboard report
Semantic Segmentation ADE20K SeMask (SeMask Swin-S FPN) Validation mIoU 47.63 #159 of 235 Archive leaderboard report
Semantic Segmentation ADE20K SeMask (SeMask Swin-T FPN) Params (M) 35 #211 of 235 Archive leaderboard report
Semantic Segmentation ADE20K SeMask (SeMask Swin-T FPN) Validation mIoU 43.16 #211 of 235 Archive leaderboard report
Semantic Segmentation ADE20K val SeMask (SeMask Swin-L FaPN-Mask2Former) mIoU 58.2 #15 of 95 Archive leaderboard report
Semantic Segmentation ADE20K val SeMask (SeMask Swin-L MSFaPN-Mask2Former) mIoU 58.2 #16 of 95 Archive leaderboard report
Semantic Segmentation ADE20K val SeMask (SeMask Swin-L Mask2Former) mIoU 57.5 #20 of 95 Archive leaderboard report
Semantic Segmentation ADE20K val SeMask (SeMask Swin-L MSFaPN-Mask2Former, single-scale) mIoU 57.0 #23 of 95 Archive leaderboard report
Semantic Segmentation ADE20K val SeMask (SeMask Swin-L MaskFormer) mIoU 56.2 #26 of 95 Archive leaderboard report
Semantic Segmentation ADE20K val SeMask (SeMask Swin-L FPN) mIoU 53.5 #40 of 95 Archive leaderboard report
Semantic Segmentation Cityscapes val SeMask (SeMask Swin-L Mask2Former) mIoU 84.98 #11 of 99 Archive leaderboard report
Semantic Segmentation Cityscapes val SeMask (SeMask Swin-L FPN) mIoU 80.39 #51 of 99 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

Absolute Position EncodingsAdamAttentionBPEDense ConnectionsDropoutLabel SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSoftmaxStochastic DepthSwin TransformerTransformer

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