Weakly supervised training of universal visual concepts for multi-domain semantic segmentation

Abstract

Deep supervised models have an unprecedented capacity to absorb large quantities of training data. Hence, training on multiple datasets becomes a method of choice towards strong generalization in usual scenes and graceful performance degradation in edge cases. Unfortunately, different datasets often have incompatible labels. For instance, the Cityscapes road class subsumes all driving surfaces, while Vistas defines separate classes for road markings, manholes etc. Furthermore, many datasets have overlapping labels. For instance, pickups are labeled as trucks in VIPER, cars in Vistas, and vans in ADE20k. We address this challenge by considering labels as unions of universal visual concepts. This allows seamless and principled learning on multi-domain dataset collections without requiring any relabeling effort. Our method achieves competitive within-dataset and cross-dataset generalization, as well as ability to learn visual concepts which are not separately labeled in any of the training datasets. Experiments reveal competitive or state-of-the-art performance on two multi-domain dataset collections and on the WildDash 2 benchmark.

arXiv comment27 pages, 16 figures, 10 tables, accepted to International Journal of Computer Vision

References in corpus (6)

• Multi-Scale Context Aggregation by Dilated Convolutions
• Per-Pixel Classification is Not All You Need for Semantic Segmentation
• The Fishyscapes Benchmark: Measuring Blind Spots in Semantic Segmentation
• SegmentMeIfYouCan: A Benchmark for Anomaly Segmentation
• Dense outlier detection and open-set recognition based on training with noisy negative images
• 1st Place Solution of The Robust Vision Challenge 2022 Semantic Segmentation Track

Cited by in corpus (1)

• Dynamic loss balancing and sequential enhancement for road-safety assessment and traffic scene classification