TasksSotADatasetsPapersMethodsSubmitAbout
Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable BenchmarksAll SotADatasetsPapersMethods

Community

Submit ResultsAbout

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Papers/Dense FixMatch: a simple semi-supervised learning method f...

Dense FixMatch: a simple semi-supervised learning method for pixel-wise prediction tasks

Miquel Martí i Rabadán, Alessandro Pieropan, Hossein Azizpour, Atsuto Maki

2022-10-18Semi-Supervised Semantic SegmentationData AugmentationSemantic Segmentation
PaperPDFCode(official)

Abstract

We propose Dense FixMatch, a simple method for online semi-supervised learning of dense and structured prediction tasks combining pseudo-labeling and consistency regularization via strong data augmentation. We enable the application of FixMatch in semi-supervised learning problems beyond image classification by adding a matching operation on the pseudo-labels. This allows us to still use the full strength of data augmentation pipelines, including geometric transformations. We evaluate it on semi-supervised semantic segmentation on Cityscapes and Pascal VOC with different percentages of labeled data and ablate design choices and hyper-parameters. Dense FixMatch significantly improves results compared to supervised learning using only labeled data, approaching its performance with 1/4 of the labeled samples.

Results

TaskDatasetMetricValueModel
Semantic SegmentationPascal VOC 2012 6.25% labeledValidation mIoU54.85Dense FixMatch (DeepLabv3+ ResNet-101, over-sampling, single pass eval)
Semantic SegmentationPascal VOC 2012 6.25% labeledValidation mIoU52.15Dense FixMatch (DeepLabv3+ ResNet-50, over-sampling, single pass eval)
Semantic SegmentationPASCAL VOC 2012 25% labeledValidation mIoU72.04Dense FixMatch (DeepLabv3+ ResNet-101, over-sampling, single pass eval)
Semantic SegmentationPASCAL VOC 2012 25% labeledValidation mIoU69.02Dense FixMatch (DeepLabv3+ ResNet-50, over-sampling, single pass eval)
Semantic SegmentationCityscapes 93 labeledValidation mIoU66.97Dense FixMatch (DeepLabv3+ ResNet-101, over-sampling, single pass eval)
Semantic SegmentationCityscapes 93 labeledValidation mIoU65.81Dense FixMatch (DeepLabv3+ ResNet-50, over-sampling, single-pass eval)
Semantic SegmentationCityscapes with extra (no coarse labels)Validation mIoU80.82Dense FixMatch (DeepLabv3+ ResNet-101, over-sampling, single pass eval)
Semantic SegmentationCityscapes with extra (no coarse labels)Validation mIoU79.98Dense FixMatch (DeepLabv3+ ResNet-50, over-sampling, single pass eval)
10-shot image generationPascal VOC 2012 6.25% labeledValidation mIoU54.85Dense FixMatch (DeepLabv3+ ResNet-101, over-sampling, single pass eval)
10-shot image generationPascal VOC 2012 6.25% labeledValidation mIoU52.15Dense FixMatch (DeepLabv3+ ResNet-50, over-sampling, single pass eval)
10-shot image generationPASCAL VOC 2012 25% labeledValidation mIoU72.04Dense FixMatch (DeepLabv3+ ResNet-101, over-sampling, single pass eval)
10-shot image generationPASCAL VOC 2012 25% labeledValidation mIoU69.02Dense FixMatch (DeepLabv3+ ResNet-50, over-sampling, single pass eval)
10-shot image generationCityscapes 93 labeledValidation mIoU66.97Dense FixMatch (DeepLabv3+ ResNet-101, over-sampling, single pass eval)
10-shot image generationCityscapes 93 labeledValidation mIoU65.81Dense FixMatch (DeepLabv3+ ResNet-50, over-sampling, single-pass eval)
10-shot image generationCityscapes with extra (no coarse labels)Validation mIoU80.82Dense FixMatch (DeepLabv3+ ResNet-101, over-sampling, single pass eval)
10-shot image generationCityscapes with extra (no coarse labels)Validation mIoU79.98Dense FixMatch (DeepLabv3+ ResNet-50, over-sampling, single pass eval)

Related Papers

SeC: Advancing Complex Video Object Segmentation via Progressive Concept Construction2025-07-21Overview of the TalentCLEF 2025: Skill and Job Title Intelligence for Human Capital Management2025-07-17Pixel Perfect MegaMed: A Megapixel-Scale Vision-Language Foundation Model for Generating High Resolution Medical Images2025-07-17DiffOSeg: Omni Medical Image Segmentation via Multi-Expert Collaboration Diffusion Model2025-07-17SCORE: Scene Context Matters in Open-Vocabulary Remote Sensing Instance Segmentation2025-07-17Unified Medical Image Segmentation with State Space Modeling Snake2025-07-17A Privacy-Preserving Semantic-Segmentation Method Using Domain-Adaptation Technique2025-07-17SAMST: A Transformer framework based on SAM pseudo label filtering for remote sensing semi-supervised semantic segmentation2025-07-16