Differential Alignment for Domain Adaptive Object Detection

TL;DR

This paper introduces a novel differential feature alignment strategy for domain adaptive object detection, focusing on instance-specific weighting and region-of-interest guidance to improve cross-domain detection performance.

Contribution

It proposes a prediction-discrepancy feedback instance alignment module and an uncertainty-based foreground-oriented image alignment module, advancing domain adaptation techniques in object detection.

Findings

Outperforms state-of-the-art methods on DAOD benchmarks.

Effective handling of domain-specific information through weighted instance alignment.

Improved focus on regions of interest enhances detection accuracy.

Abstract

Domain adaptive object detection (DAOD) aims to generalize an object detector trained on labeled source-domain data to a target domain without annotations, the core principle of which is \emph{source-target feature alignment}. Typically, existing approaches employ adversarial learning to align the distributions of the source and target domains as a whole, barely considering the varying significance of distinct regions, say instances under different circumstances and foreground \emph{vs} background areas, during feature alignment. To overcome the shortcoming, we investigates a differential feature alignment strategy. Specifically, a prediction-discrepancy feedback instance alignment module (dubbed PDFA) is designed to adaptively assign higher weights to instances of higher teacher-student detection discrepancy, effectively handling heavier domain-specific information. Additionally, an uncertainty-based foreground-oriented image alignment module (UFOA) is proposed to explicitly guide the model to focus more on regions of interest. Extensive experiments on widely-used DAOD datasets together with ablation studies are conducted to demonstrate the efficacy of our proposed method and reveal its superiority over other SOTA alternatives. Our code is available at https://github.com/EstrellaXyu/Differential-Alignment-for-DAOD.