Significance-aware Information Bottleneck for Domain Adaptive Semantic Segmentation

Abstract

For unsupervised domain adaptation problems, the strategy of aligning the two domains in latent feature space through adversarial learning has achieved much progress in image classification, but usually fails in semantic segmentation tasks in which the latent representations are overcomplex. In this work, we equip the adversarial network with a "significance-aware information bottleneck (SIB)", to address the above problem. The new network structure, called SIBAN, enables a significance-aware feature purification before the adversarial adaptation, which eases the feature alignment and stabilizes the adversarial training course. In two domain adaptation tasks, i.e., GTA5 -> Cityscapes and SYNTHIA -> Cityscapes, we validate that the proposed method can yield leading results compared with other feature-space alternatives. Moreover, SIBAN can even match the state-of-the-art output-space methods in segmentation accuracy, while the latter are often considered to be better choices for domain adaptive segmentation task.

arXiv comment12 pages, 9 figures

References in corpus (19)

• Adam: A Method for Stochastic Optimization
• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Deep Residual Learning for Image Recognition
• The information bottleneck method
• Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks
• Progressive Growing of GANs for Improved Quality, Stability, and Variation
• The Cityscapes Dataset for Semantic Urban Scene Understanding
• Fully Convolutional Networks for Semantic Segmentation
• FCNs in the Wild: Pixel-level Adversarial and Constraint-based Adaptation
• CyCADA: Cycle-Consistent Adversarial Domain Adaptation
• Adversarial Discriminative Domain Adaptation
• Deep Variational Information Bottleneck
• Dilated Residual Networks
• Playing for Data: Ground Truth from Computer Games
• Maximum Classifier Discrepancy for Unsupervised Domain Adaptation
• Unsupervised Pixel-Level Domain Adaptation with Generative Adversarial Networks
• Semantic-aware Grad-GAN for Virtual-to-Real Urban Scene Adaption
• Relevant sparse codes with variational information bottleneck

Cited by in corpus (6)

• Adversarial Style Mining for One-Shot Unsupervised Domain Adaptation
• Domain adaptation based self-correction model for COVID-19 infection segmentation in CT images
• Learning Texture Invariant Representation for Domain Adaptation of Semantic Segmentation
• Rectifying Pseudo Label Learning via Uncertainty Estimation for Domain Adaptive Semantic Segmentation
• Contextual-Relation Consistent Domain Adaptation for Semantic Segmentation
• Context-Aware Domain Adaptation in Semantic Segmentation