Robust Semantic Segmentation with Superpixel-Mix
This work addresses reliability issues (robustness, uncertainty, bias) in semantic segmentation for real-world applications, representing an incremental improvement over existing mixing-based augmentation techniques.
The paper tackles the problem of improving reliability in semantic segmentation by introducing Superpixel-mix, a data augmentation method that mixes superpixels between images with teacher-student consistency training. It achieves state-of-the-art results on Cityscapes for semi-supervised segmentation and shows competitive performance under distribution shifts like adverse weather and image corruptions.
Along with predictive performance and runtime speed, reliability is a key requirement for real-world semantic segmentation. Reliability encompasses robustness, predictive uncertainty and reduced bias. To improve reliability, we introduce Superpixel-mix, a new superpixel-based data augmentation method with teacher-student consistency training. Unlike other mixing-based augmentation techniques, mixing superpixels between images is aware of object boundaries, while yielding consistent gains in segmentation accuracy. Our proposed technique achieves state-of-the-art results in semi-supervised semantic segmentation on the Cityscapes dataset. Moreover, Superpixel-mix improves the reliability of semantic segmentation by reducing network uncertainty and bias, as confirmed by competitive results under strong distributions shift (adverse weather, image corruptions) and when facing out-of-distribution data.