← Search

Christoph Kamann

2 accepted papers

2020

Increasing the Robustness of Semantic Segmentation Models with Painting-by-Numbers

ECCV 2020poster

For safety-critical applications such as autonomous driving, CNNs have to be robust with respect to unavoidable image corruptions, such as image noise. While previous works addressed the task of robust prediction in the context of full-image classification, we consider it for dense semantic segmenta…

Cited by 26SourcePDFScholar