Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic Segmentation (2022-08-21T00:00:00.000000Z)

TL;DR

This work revisits the weak-to-strong consistency framework, popularized by FixMatch from semi-supervised classification, and presents a dual-stream perturbation technique, enabling two strong views to be simultaneously guided by a common weak view.

Abstract

In this work, we revisit the weak-to-strong consistency framework, popularized by FixMatch from semi-supervised classification, where the prediction of a weakly perturbed image serves as supervision for its strongly perturbed version. Intriguingly, we observe that such a simple pipeline already achieves competitive results against recent advanced works, when transferred to our segmentation scenario. Its success heavily relies on the manual design of strong data augmentations, however, which may be limited and inadequate to explore a broader perturbation space. Motivated by this, we propose an auxiliary feature perturbation stream as a supplement, leading to an expanded perturbation space. On the other, to sufficiently probe original image-level augmentations, we present a dual-stream perturbation technique, enabling two strong views to be simultaneously guided by a common weak view. Consequently, our overall Unified Dual-Stream Perturbations approach (UniMatch) surpasses all existing methods significantly across all evaluation protocols on the Pascal, Cityscapes, and COCO benchmarks. Its superiority is also demonstrated in remote sensing interpretation and medical image analysis. We hope our reproduced FixMatch and our results can inspire more future works.

Authors

Wayne Zhang

11 papers

Lei Qi

3 papers

Litong Feng

4 papers

TL;DR

Abstract

Authors

References82 items

Semi-supervised Semantic Segmentation with Prototype-based Consistency Regularization

FreeMatch: Self-adaptive Thresholding for Semi-supervised Learning

Region-level Contrastive and Consistency Learning for Semi-Supervised Semantic Segmentation

Revisiting Consistency Regularization for Semi-supervised Change Detection in Remote Sensing Images

Semi-supervised Semantic Segmentation with Error Localization Network

Unbiased Subclass Regularization for Semi-Supervised Semantic Segmentation

Semi-Supervised Semantic Segmentation Using Unreliable Pseudo-Labels

Cross-Model Pseudo-Labeling for Semi-Supervised Action Recognition

Generative Adversarial Networks

Semi-Supervised Medical Image Segmentation via Cross Teaching between CNN and Transformer

Perturbed and Strict Mean Teachers for Semi-supervised Semantic Segmentation

Learning from Temporal Gradient for Semi-supervised Action Recognition

FlexMatch: Boosting Semi-Supervised Learning with Curriculum Pseudo Labeling

Semi-Supervised Semantic Segmentation via Adaptive Equalization Learning

C3-SemiSeg: Contrastive Semi-supervised Segmentation via Cross-set Learning and Dynamic Class-balancing

Pixel Contrastive-Consistent Semi-Supervised Semantic Segmentation

Re-distributing Biased Pseudo Labels for Semi-supervised Semantic Segmentation: A Baseline Investigation

The Bottleneck

End-to-End Semi-Supervised Object Detection with Soft Teacher

ST++: Make Self-trainingWork Better for Semi-supervised Semantic Segmentation

ATSO: Asynchronous Teacher-Student Optimization for Semi-Supervised Image Segmentation

Semi-Supervised Semantic Segmentation with Cross Pseudo Supervision

Semi-supervised Semantic Segmentation with Directional Context-aware Consistency

Humble Teachers Teach Better Students for Semi-Supervised Object Detection

Robust Mutual Learning for Semi-supervised Semantic Segmentation

PixMatch: Unsupervised Domain Adaptation via Pixelwise Consistency Training

Semi-Supervised Semantic Segmentation with Pixel-Level Contrastive Learning from a Class-wise Memory Bank

A Simple Baseline for Semi-supervised Semantic Segmentation with Strong Data Augmentation*

Bootstrapping Semantic Segmentation with Regional Contrast

Unbiased Teacher for Semi-Supervised Object Detection

CoMatch: Semi-supervised Learning with Contrastive Graph Regularization

AlphaMatch: Improving Consistency for Semi-supervised Learning with Alpha-divergence

Exploring Simple Siamese Representation Learning

PseudoSeg: Designing Pseudo Labels for Semantic Segmentation

Guided Collaborative Training for Pixel-wise Semi-Supervised Learning

SemiCDNet: A Semisupervised Convolutional Neural Network for Change Detection in High Resolution Remote-Sensing Images

FeatMatch: Feature-Based Augmentation for Semi-Supervised Learning

Unsupervised Learning of Visual Features by Contrasting Cluster Assignments

Rethinking Pre-training and Self-training

A Spatial-Temporal Attention-Based Method and a New Dataset for Remote Sensing Image Change Detection

DMT: Dynamic mutual training for semi-supervised learning

Meta Pseudo Labels

Semi-Supervised Semantic Segmentation With Cross-Consistency Training

A Simple Framework for Contrastive Learning of Visual Representations

FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence

ReMixMatch: Semi-Supervised Learning with Distribution Alignment and Augmentation Anchoring

Momentum Contrast for Unsupervised Visual Representation Learning

Self-Training With Noisy Student Improves ImageNet Classification

Semi-supervised semantic segmentation needs strong, high-dimensional perturbations

Semi-Supervised Semantic Segmentation With High- and Low-Level Consistency

Pseudo-Labeling and Confirmation Bias in Deep Semi-Supervised Learning

Uncertainty-aware Self-ensembling Model for Semi-supervised 3D Left Atrium Segmentation

CutMix: Regularization Strategy to Train Strong Classifiers With Localizable Features

MixMatch: A Holistic Approach to Semi-Supervised Learning

Unsupervised Data Augmentation for Consistency Training

Fully Convolutional Networks for Multisource Building Extraction From an Open Aerial and Satellite Imagery Data Set

ADVENT: Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentation

Representation Learning with Contrastive Predictive Coding

Deep Learning Techniques for Automatic MRI Cardiac Multi-Structures Segmentation and Diagnosis: Is the Problem Solved?

Context Encoding for Semantic Segmentation

Deep Co-Training for Semi-Supervised Image Recognition

Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation

Semi Supervised Semantic Segmentation Using Generative Adversarial Network

Improved Regularization of Convolutional Neural Networks with Cutout

Deep Mutual Learning

Virtual Adversarial Training: A Regularization Method for Supervised and Semi-Supervised Learning

Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results

Pyramid Scene Parsing Network

Xception: Deep Learning with Depthwise Separable Convolutions

Temporal Ensembling for Semi-Supervised Learning

Regularization With Stochastic Transformations and Perturbations for Deep Semi-Supervised Learning

The Cityscapes Dataset for Semantic Urban Scene Understanding

Deep Residual Learning for Image Recognition

Fully convolutional networks for semantic segmentation

The Pascal Visual Object Classes Challenge: A Retrospective

Microsoft COCO: Common Objects in Context