Semi-Supervised Semantic Segmentation via Boosting Uncertainty on Unlabeled Data

Daoan Zhang, Yunhao Luo, Jie-Neng Chen, Lingyun Huang, Jianguo Zhang

Keywords: Semantic Segmentation, Semi-supervised Learning, Uncertainty in Deep Learning

TL;DR: We theoretically analyze and experimentally prove that appropriately boosting uncertainty on unlabeled data can help minimize the distribution gap in semi-supervised semantic segmentation.

Abstract: We bring a new perspective to semi-supervised semantic segmentation by providing an analysis on the labeled and unlabeled distributions in training datasets. We firstly figure out that the distribution gap between labeled and unlabeled datasets cannot be ignored, even though the two datasets are sampled from the same distribution. To address this issue, we theoretically analyze and experimentally prove that appropriately boosting uncertainty on unlabeled data can help minimize the distribution gap, which benefits the generalization of the model. We propose two strategies and design an algorithm of uncertainty booster specially for semi-supervised semantic segmentation. Extensive experiments are carried out based on these theories, and the results confirmed the efficacy of the algorithm and strategies. Our plug-and play uncertainty booster is tiny, efficient and robust to hyper parameters, but can significantly promote the performance. In our experiments, our method achieves state-of-the-art performance compared to the current semi-supervised semantic segmentation methods on the popular benchmark: Cityscapes and PASCAL VOC 2012 with different train settings.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Deep Learning and representational learning