Learning to adapt class-specific features across domains for semantic segmentation

TL;DR

This paper introduces a class-specific feature adaptation method for semantic segmentation across domains, utilizing a conditional discriminator and StarGAN for improved detail preservation without extensive domain-specific parameters.

Contribution

It proposes a novel architecture that adapts features per class using a conditional pixel-wise discriminator and leverages StarGAN for multi-domain image translation.

Findings

Improves segmentation accuracy over strong baselines

Effectively preserves details in domain adaptation

Demonstrates potential in preliminary experiments

Abstract

Recent advances in unsupervised domain adaptation have shown the effectiveness of adversarial training to adapt features across domains, endowing neural networks with the capability of being tested on a target domain without requiring any training annotations in this domain. The great majority of existing domain adaptation models rely on image translation networks, which often contain a huge amount of domain-specific parameters. Additionally, the feature adaptation step often happens globally, at a coarse level, hindering its applicability to tasks such as semantic segmentation, where details are of crucial importance to provide sharp results. In this thesis, we present a novel architecture, which learns to adapt features across domains by taking into account per class information. To that aim, we design a conditional pixel-wise discriminator network, whose output is conditioned on the segmentation masks. Moreover, following recent advances in image translation, we adopt the recently introduced StarGAN architecture as image translation backbone, since it is able to perform translations across multiple domains by means of a single generator network. Preliminary results on a segmentation task designed to assess the effectiveness of the proposed approach highlight the potential of the model, improving upon strong baselines and alternative designs.