Latent Space Regularization for Unsupervised Domain Adaptation in Semantic Segmentation
This addresses the challenge of poor generalization in semantic segmentation for autonomous driving, though it is incremental as it builds on existing domain adaptation methods.
The paper tackles the problem of domain shift in semantic segmentation for autonomous driving by introducing feature-level space-shaping regularization strategies, achieving state-of-the-art results in synthetic-to-real benchmarks.
Deep convolutional neural networks for semantic segmentation achieve outstanding accuracy, however they also have a couple of major drawbacks: first, they do not generalize well to distributions slightly different from the one of the training data; second, they require a huge amount of labeled data for their optimization. In this paper, we introduce feature-level space-shaping regularization strategies to reduce the domain discrepancy in semantic segmentation. In particular, for this purpose we jointly enforce a clustering objective, a perpendicularity constraint and a norm alignment goal on the feature vectors corresponding to source and target samples. Additionally, we propose a novel measure able to capture the relative efficacy of an adaptation strategy compared to supervised training. We verify the effectiveness of such methods in the autonomous driving setting achieving state-of-the-art results in multiple synthetic-to-real road scenes benchmarks.