To Adapt or Not to Adapt? Real-Time Adaptation for Semantic Segmentation

TL;DR

This paper introduces HAMLET, a real-time, hardware-aware framework for online domain adaptation in semantic segmentation, enabling fast adaptation during deployment with minimal computational overhead.

Contribution

We propose HAMLET, a novel hardware-aware modular training framework with active domain-shift detection for real-time semantic segmentation adaptation.

Findings

Achieves over 29FPS on a single consumer GPU.

Demonstrates effective adaptation on OnDA and SHIFT benchmarks.

Balances accuracy and speed in real-time domain adaptation.

Abstract

The goal of Online Domain Adaptation for semantic segmentation is to handle unforeseeable domain changes that occur during deployment, like sudden weather events. However, the high computational costs associated with brute-force adaptation make this paradigm unfeasible for real-world applications. In this paper we propose HAMLET, a Hardware-Aware Modular Least Expensive Training framework for real-time domain adaptation. Our approach includes a hardware-aware back-propagation orchestration agent (HAMT) and a dedicated domain-shift detector that enables active control over when and how the model is adapted (LT). Thanks to these advancements, our approach is capable of performing semantic segmentation while simultaneously adapting at more than 29FPS on a single consumer-grade GPU. Our framework's encouraging accuracy and speed trade-off is demonstrated on OnDA and SHIFT benchmarks through experimental results.