Semi-Supervised Semantic Segmentation via Marginal Contextual Information

Moshe Kimhi, Shai Kimhi, Evgenii Zheltonozhskii, Or Litany, Chaim Baskin
8/26/2023
cs.CVcs.LG

No official code found. Best implementation: s4mcontext/s4mc (9 stars, MIT, last commit 2024-05-25). 1 more below.

Abstract

We present a novel confidence refinement scheme that enhances pseudo labels in semi-supervised semantic segmentation. Unlike existing methods, which filter pixels with low-confidence predictions in isolation, our approach leverages the spatial correlation of labels in segmentation maps by grouping neighboring pixels and considering their pseudo labels collectively. With this contextual information, our method, named S4MC, increases the amount of unlabeled data used during training while maintaining the quality of the pseudo labels, all with negligible computational overhead. Through extensive experiments on standard benchmarks, we demonstrate that S4MC outperforms existing state-of-the-art semi-supervised learning approaches, offering a promising solution for reducing the cost of acquiring dense annotations. For example, S4MC achieves a 1.39 mIoU improvement over the prior art on PASCAL VOC 12 with 366 annotated images. The code to reproduce our experiments is available at https://s4mcontext.github.io/

View on arXivView PDF

Code Implementations(2)

s4mcontext/s4mcCommunity100%
MIT

code for s4mc

91PythonMar 15, 20232 years agoMIT
shaiTheKimhi/s4mcCommunity67%
MIT

Semi-Supervised Semantic Segmentation via Marginal Contextual Information

40Shell, PythonMay 30, 20242 years agoMIT

Similar Papers

When Confidence Fails: Revisiting Pseudo-Label Selection in Semi-supervised Semantic Segmentation

92% match1 repo202025

Weighting Pseudo-Labels via High-Activation Feature Index Similarity and Object Detection for Semi-Supervised Segmentation

91% match1 repo32026

CorrMatch: Label Propagation via Correlation Matching for Semi-Supervised Semantic Segmentation

90% match2 repos1632026

Accuracy Improvement of Semi-Supervised Segmentation Using Supervised ClassMix and Sup-Unsup Feature Discriminator

89% match2026

RePL: Pseudo-label Refinement for Semi-supervised LiDAR Semantic Segmentation

88% match2026

Enhanced Soft Label for Semi-Supervised Semantic Segmentation

88% match1 repo52026

Cite this paper

@article{kimhi2023semisupervised,
  title  = {Semi-Supervised Semantic Segmentation via Marginal Contextual Information},
  author = {Moshe Kimhi and Shai Kimhi and Evgenii Zheltonozhskii and Or Litany and Chaim Baskin},
  year   = {2023},
  eprint = {2308.13900},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2308.13900v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2308.13900.svg)](https://codeofpaper.com/paper/2308.13900)

Discussion