Papers › Uncertainty Estimation via Response Scaling for Pseudo-mask Noise Mitigation in...

Uncertainty Estimation via Response Scaling for Pseudo-mask Noise Mitigation in Weakly-supervised Semantic Segmentation

14 Dec 2021arXiv:2112.07431archive 2025-07-28

Yi Li, Yiqun Duan, Zhanghui Kuang, Yimin Chen, Wayne Zhang, Xiaomeng Li

Weakly-Supervised Semantic Segmentation (WSSS) segments objects without a heavy burden of dense annotation. While as a price, generated pseudo-masks exist obvious noisy pixels, which result in sub-optimal segmentation models trained over these pseudo-masks. But rare studies notice or work on this problem, even these noisy pixels are inevitable after their improvements on pseudo-mask. So we try to improve WSSS in the aspect of noise mitigation. And we observe that many noisy pixels are of high confidence, especially when the response range is too wide or narrow, presenting an uncertain status. Thus, in this paper, we simulate noisy variations of response by scaling the prediction map multiple times for uncertainty estimation. The uncertainty is then used to weight the segmentation loss to mitigate noisy supervision signals. We call this method URN, abbreviated from Uncertainty estimation via Response scaling for Noise mitigation. Experiments validate the benefits of URN, and our method achieves state-of-the-art results at 71.2% and 41.5% on PASCAL VOC 2012 and MS COCO 2014 respectively, without extra models like saliency detection. Code is available at https://github.com/XMed-Lab/URN.

PaperPDFCode

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

Code

xmed-lab/urn officialmentioned in papermentioned on GitHubpytorch 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

Saliency DetectionSegmentationSemantic SegmentationWeakly supervised Semantic SegmentationWeakly-Supervised Semantic Segmentation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Weakly-Supervised Semantic Segmentation COCO 2014 val URN(Res2Net-101, no saliency, no RW) mIoU 41.5 #24 of 39 Archive leaderboard report
Weakly-Supervised Semantic Segmentation COCO 2014 val URN(ScaleNet-101, no saliency, no RW) mIoU 40.8 #25 of 39 Archive leaderboard report
Weakly-Supervised Semantic Segmentation COCO 2014 val URN(ResNet-101, no saliency, no RW) mIoU 40.7 #26 of 39 Archive leaderboard report
Weakly-Supervised Semantic Segmentation COCO 2014 val URN(ResNet-38, no saliency, no RW) mIoU 40.5 #27 of 39 Archive leaderboard report
Weakly-Supervised Semantic Segmentation PASCAL VOC 2012 test URN(Res2Net-101, no saliency, no RW) Mean IoU 71.5 #27 of 60 Archive leaderboard report
Weakly-Supervised Semantic Segmentation PASCAL VOC 2012 test URN(ScaleNet-101, no saliency, no RW) Mean IoU 70.8 #33 of 60 Archive leaderboard report
Weakly-Supervised Semantic Segmentation PASCAL VOC 2012 test URN(ResNet-38, no saliency, no RW) Mean IoU 70.6 #38 of 60 Archive leaderboard report
Weakly-Supervised Semantic Segmentation PASCAL VOC 2012 test URN(ResNet-101, no saliency, no RW) Mean IoU 69.7 #45 of 60 Archive leaderboard report
Weakly-Supervised Semantic Segmentation PASCAL VOC 2012 val URN(Res2Net-101, no saliency, no RW) Mean IoU 71.2 #26 of 73 Archive leaderboard report
Weakly-Supervised Semantic Segmentation PASCAL VOC 2012 val URN(ScaleNet-101, no saliency, no RW) Mean IoU 70.1 #43 of 73 Archive leaderboard report
Weakly-Supervised Semantic Segmentation PASCAL VOC 2012 val URN(ResNet-101, no saliency, no RW) Mean IoU 69.5 #46 of 73 Archive leaderboard report
Weakly-Supervised Semantic Segmentation PASCAL VOC 2012 val URN(ResNet-38, no saliency, no RW) Mean IoU 69.4 #50 of 73 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.

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