Weakly Supervised Semantic Segmentation via Alternate Self-Dual Teaching

Dingwen Zhang, Hao Li, Wenyuan Zeng, Chaowei Fang, Lechao Cheng, Ming Ming Cheng, Junwei Han

Research output: Contribution to journalArticlepeer-review

35 Scopus citations

Abstract

Weakly supervised semantic segmentation (WSSS) is a challenging yet important research field in vision community. In WSSS, the key problem is to generate high-quality pseudo segmentation masks (PSMs). Existing approaches mainly depend on the discriminative object part to generate PSMs, which would inevitably miss object parts or involve surrounding image background, as the learning process is unaware of the full object structure. In fact, both the discriminative object part and the full object structure are critical for deriving of high-quality PSMs. To fully explore these two information cues, we build a novel end-to-end learning framework, alternate self-dual teaching (ASDT), based on a dual-teacher single-student network architecture. The information interaction among different network branches is formulated in the form of knowledge distillation (KD). Unlike the conventional KD, the knowledge of the two teacher models would inevitably be noisy under weak supervision. Inspired by the Pulse Width (PW) modulation, we introduce a PW wave-like selection signal to alleviate the influence of the imperfect knowledge from either teacher model on the KD process. Comprehensive experiments on the PASCAL VOC 2012 and COCO-Stuff 10K demonstrate the effectiveness of the proposed ASDT framework, and new state-of-the-art results are achieved.

Original languageEnglish
Pages (from-to)3086-3095
Number of pages10
JournalIEEE Transactions on Image Processing
Volume34
DOIs
StatePublished - 2025

Keywords

  • Weakly supervised learning
  • dual teaching
  • knowledge distillation
  • semantic segmentation

Fingerprint

Dive into the research topics of 'Weakly Supervised Semantic Segmentation via Alternate Self-Dual Teaching'. Together they form a unique fingerprint.

Cite this