Rethinking Ensemble-Distillation for Semantic Segmentation Based Unsupervised Domain Adaptation

TL;DR

This paper introduces a flexible ensemble-distillation framework for semantic segmentation in unsupervised domain adaptation, enabling arbitrary ensemble compositions while maintaining high performance and robustness across benchmarks.

Contribution

The paper proposes a novel, flexible ensemble-distillation framework that allows arbitrary ensemble compositions for semantic segmentation UDA, addressing rigidity issues of previous end-to-end methods.

Findings

Achieves superior performance on GTA5 to Cityscapes and SYNTHIA to Cityscapes benchmarks.

Demonstrates robustness against output inconsistency and member performance variation.

Provides detailed analysis validating the effectiveness of the framework.

Abstract

Recent researches on unsupervised domain adaptation (UDA) have demonstrated that end-to-end ensemble learning frameworks serve as a compelling option for UDA tasks. Nevertheless, these end-to-end ensemble learning methods often lack flexibility as any modification to the ensemble requires retraining of their frameworks. To address this problem, we propose a flexible ensemble-distillation framework for performing semantic segmentation based UDA, allowing any arbitrary composition of the members in the ensemble while still maintaining its superior performance. To achieve such flexibility, our framework is designed to be robust against the output inconsistency and the performance variation of the members within the ensemble. To examine the effectiveness and the robustness of our method, we perform an extensive set of experiments on both GTA5 to Cityscapes and SYNTHIA to Cityscapes benchmarks to quantitatively inspect the improvements achievable by our method. We further provide detailed analyses to validate that our design choices are practical and beneficial. The experimental evidence validates that the proposed method indeed offer superior performance, robustness and flexibility in semantic segmentation based UDA tasks against contemporary baseline methods.