Re-distributing Biased Pseudo Labels for Semi-supervised Semantic Segmentation: A Baseline Investigation
This addresses the issue of class imbalance in semi-supervised semantic segmentation for real-world applications, but it is incremental as it builds on existing self-training methods.
The paper tackles the problem of biased pseudo labels in semi-supervised semantic segmentation due to long-tailed class distributions, proposing a method that produces unbiased pseudo labels and achieves favorable performance compared to state-of-the-art approaches on Cityscapes and PASCAL VOC 2012 datasets.
While self-training has advanced semi-supervised semantic segmentation, it severely suffers from the long-tailed class distribution on real-world semantic segmentation datasets that make the pseudo-labeled data bias toward majority classes. In this paper, we present a simple and yet effective Distribution Alignment and Random Sampling (DARS) method to produce unbiased pseudo labels that match the true class distribution estimated from the labeled data. Besides, we also contribute a progressive data augmentation and labeling strategy to facilitate model training with pseudo-labeled data. Experiments on both Cityscapes and PASCAL VOC 2012 datasets demonstrate the effectiveness of our approach. Albeit simple, our method performs favorably in comparison with state-of-the-art approaches. Code will be available at https://github.com/CVMI-Lab/DARS.