Attention-based Domain Adaptation for Single Stage Detectors

TL;DR

This paper introduces an attention-based domain adaptation method for single-stage object detectors, enhancing their performance across different data distributions by focusing on important regions without region proposals.

Contribution

The authors propose a novel attention mechanism that enables local adaptation in single-stage detectors, outperforming existing methods designed for specific architectures.

Findings

Outperforms state-of-the-art domain adaptation techniques on benchmark datasets.

Effectively adapts features from global to local levels in single-stage detectors.

Applicable to detectors like SSD and YOLOv5 with improved accuracy.

Abstract

While domain adaptation has been used to improve the performance of object detectors when the training and test data follow different distributions, previous work has mostly focused on two-stage detectors. This is because their use of region proposals makes it possible to perform local adaptation, which has been shown to significantly improve the adaptation effectiveness. Here, by contrast, we target single-stage architectures, which are better suited to resource-constrained detection than two-stage ones but do not provide region proposals. To nonetheless benefit from the strength of local adaptation, we introduce an attention mechanism that lets us identify the important regions on which adaptation should focus. Our method gradually adapts the features from global, image-level to local, instance-level. Our approach is generic and can be integrated into any single-stage detector. We demonstrate this on standard benchmark datasets by applying it to both SSD and YOLOv5. Furthermore, for equivalent single-stage architectures, our method outperforms the state-of-the-art domain adaptation techniques even though they were designed for specific detectors.