Robust Semantic Segmentation with Superpixel-Mix

Gianni Franchi   Nacim Belkhir   Mai Lan Ha   Yufei Hu   Andrei Bursuc   Volker Blanz   Angela Yao

BMVC 2021

Abstract

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.

BibTeX

@inproceedings{franchi2021robust,
  title={Robust Semantic Segmentation with Superpixel-Mix},
  author={Franchi, Gianni and Belkhir, Nacim and Ha, Mai Lan and Hu, Yufei and Bursuc, Andrei and Blanz, Volker and Yao, Angela},
  booktitle={BMVC},
  year={2021}
}