Adaptive Early-Learning Correction for Segmentation from Noisy Annotations
arXiv:2110.03740
Abstract
Deep learning in the presence of noisy annotations has been studied extensively in classification, but much less in segmentation tasks. In this work, we study the learning dynamics of deep segmentation networks trained on inaccurately-annotated data. We discover a phenomenon that has been previously reported in the context of classification: the networks tend to first fit the clean pixel-level labels during an "early-learning" phase, before eventually memorizing the false annotations. However, in contrast to classification, memorization in segmentation does not arise simultaneously for all semantic categories. Inspired by these findings, we propose a new method for segmentation from noisy annotations with two key elements. First, we detect the beginning of the memorization phase separately for each category during training. This allows us to adaptively correct the noisy annotations in order to exploit early learning. Second, we incorporate a regularization term that enforces consistency across scales to boost robustness against annotation noise. Our method outperforms standard approaches on a medical-imaging segmentation task where noises are synthesized to mimic human annotation errors. It also provides robustness to realistic noisy annotations present in weakly-supervised semantic segmentation, achieving state-of-the-art results on PASCAL VOC 2012. Code is available at https://github.com/Kangningthu/ADELE
The first two authors contribute equally, order decided by coin flipping. Accepted by CVPR 2022
References in corpus (8)
- Efficient Inference in Fully Connected CRFs with Gaussian Edge Potentials
- FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
- Training Deep Neural Networks on Noisy Labels with Bootstrapping
- A Closer Look at Memorization in Deep Networks
- Disentangling Human Error from the Ground Truth in Segmentation of Medical Images
- Structured Consistency Loss for semi-supervised semantic segmentation
- Non-Salient Region Object Mining for Weakly Supervised Semantic Segmentation
- Weakly-supervised High-resolution Segmentation of Mammography Images for Breast Cancer Diagnosis