Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Weakly Supervised Co-training with Swapping Assignments for Semantic Segmentation

About

Class activation maps (CAMs) are commonly employed in weakly supervised semantic segmentation (WSSS) to produce pseudo-labels. Due to incomplete or excessive class activation, existing studies often resort to offline CAM refinement, introducing additional stages or proposing offline modules. This can cause optimization difficulties for single-stage methods and limit generalizability. In this study, we aim to reduce the observed CAM inconsistency and error to mitigate reliance on refinement processes. We propose an end-to-end WSSS model incorporating guided CAMs, wherein our segmentation model is trained while concurrently optimizing CAMs online. Our method, Co-training with Swapping Assignments (CoSA), leverages a dual-stream framework, where one sub-network learns from the swapped assignments generated by the other. We introduce three techniques: i) soft perplexity-based regularization to penalize uncertain regions; ii) a threshold-searching approach to dynamically revise the confidence threshold; and iii) contrastive separation to address the coexistence problem. CoSA demonstrates exceptional performance, achieving mIoU of 76.2\% and 51.0\% on VOC and COCO validation datasets, respectively, surpassing existing baselines by a substantial margin. Notably, CoSA is the first single-stage approach to outperform all existing multi-stage methods including those with additional supervision. Code is avilable at \url{https://github.com/youshyee/CoSA}.

Xinyu Yang, Hossein Rahmani, Sue Black, Bryan M. Williams• 2024

Related benchmarks

TaskDatasetResultRank
Semantic segmentationPASCAL VOC 2012 (val)
mIoU81.4
166
Semantic segmentationBHSD IPH
Dice46.9
18
Semantic segmentationBHSD SDH
Dice Coefficient55.1
18
Weakly supervised semantic segmentationBraTS Tumor Core 2020 (val)
Dice Score67.7
18
Weakly supervised semantic segmentationBraTS Edema 2020 (val)
Dice47.6
18
Semantic segmentationBHSD EPH
Dice47.6
18
Semantic segmentationBHSD IVH
Dice Coefficient46.6
18
Showing 7 of 7 rows

Other info

Follow for update