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

TL;DR

This paper introduces a method for training universal visual concepts across multiple datasets for semantic segmentation, enabling better generalization and concept learning without relabeling.

Contribution

It proposes a label-union approach that allows seamless multi-domain training and concept learning without relabeling efforts.

Findings

Achieves state-of-the-art results on multi-domain datasets.

Demonstrates improved cross-dataset generalization.

Learns visual concepts not explicitly labeled in datasets.

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.