PseudoSeg: Designing Pseudo Labels for Semantic Segmentation

Abstract

Recent advances in semi-supervised learning (SSL) demonstrate that a combination of consistency regularization and pseudo-labeling can effectively improve image classification accuracy in the low-data regime. Compared to classification, semantic segmentation tasks require much more intensive labeling costs. Thus, these tasks greatly benefit from data-efficient training methods. However, structured outputs in segmentation render particular difficulties (e.g., designing pseudo-labeling and augmentation) to apply existing SSL strategies. To address this problem, we present a simple and novel re-design of pseudo-labeling to generate well-calibrated structured pseudo labels for training with unlabeled or weakly-labeled data. Our proposed pseudo-labeling strategy is network structure agnostic to apply in a one-stage consistency training framework. We demonstrate the effectiveness of the proposed pseudo-labeling strategy in both low-data and high-data regimes. Extensive experiments have validated that pseudo labels generated from wisely fusing diverse sources and strong data augmentation are crucial to consistency training for segmentation. The source code is available at https://github.com/googleinterns/wss.

arXiv commentICLR 2021. Project page: https://yuliang.vision/pseudo_seg/ Code: https://github.com/googleinterns/wss

References in corpus (10)

• Efficient Inference in Fully Connected CRFs with Gaussian Edge Potentials
• Improved Regularization of Convolutional Neural Networks with Cutout
• On Calibration of Modern Neural Networks
• Unsupervised Data Augmentation for Consistency Training
• Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results
• Rethinking Pre-training and Self-training
• A Simple Semi-Supervised Learning Framework for Object Detection
• Virtual Adversarial Training: A Regularization Method for Supervised and Semi-Supervised Learning
• Structured Consistency Loss for semi-supervised semantic segmentation
• Improving Semantic Segmentation via Self-Training

Cited by in corpus (14)

• Learning Self-Supervised Low-Rank Network for Single-Stage Weakly and Semi-Supervised Semantic Segmentation
• Reference-guided Pseudo-Label Generation for Medical Semantic Segmentation
• ST++: Make Self-training Work Better for Semi-supervised Semantic Segmentation
• Perturbed and Strict Mean Teachers for Semi-supervised Semantic Segmentation
• Semi-Supervised Semantic Segmentation via Adaptive Equalization Learning
• Bootstrap Representation Learning for Segmentation on Medical Volumes and Sequences
• Robust Mutual Learning for Semi-supervised Semantic Segmentation
• Learning from Temporal Gradient for Semi-supervised Action Recognition
• Taking it further: leveraging pseudo labels for field delineation across label-scarce smallholder regions
• SegMatch: A semi-supervised learning method for surgical instrument segmentation
• MS-KD: Multi-Organ Segmentation with Multiple Binary-Labeled Datasets
• A Simple Baseline for Semi-supervised Semantic Segmentation with Strong Data Augmentation
• Weakly Supervised Pseudo-Label assisted Learning for ALS Point Cloud Semantic Segmentation
• Vessel-CAPTCHA: an efficient learning framework for vessel annotation and segmentation