Mixed Global and Local Attention Alleviates Domain Shift Between Terahertz Image Datasets

Terahertz imaging technology has a broad application prospect in the field of security inspection due to its strong penetrability and negligible radiation effect on people. With the development of machine learning in recent years, the application of machine learning methods to security inspection can save labor costs and ensure the accuracy of long-time security inspection. However, terahertz imagers produce different bottom noise at different temperatures, which leads to a significant domain offset problem between data sets collected at different temperatures. This domain shift makes the model trained on the source domain dataset to have a significant error when applied to the target domain dataset. Therefore, how to overcome the temperature-induced domain offset is an urgent problem for terahertz imaging techniques. In this paper, we innovatively propose to apply the attention mechanism to solve the problem of domain bias in object detection. We propose a lightweight attention mechanism Mixed global and local attention(MGLA) aimed at mitigating the domain shift between the target and source domains without affecting the source domain detection effect. MGLA can take into account the local information while fusing the channel information and spatial information and global information to improve the representation of the network. The experiment results show a 10% improvement in mAP0.5 and a 6.3% improvement in mAP50:95 compared to YOLOV8 baseline.

I. Introduction

Terahertz (THz) imaging technology, as a cutting-edge means of security inspection, has attracted global attention for its superior performance of harmlessness and diversity of detection. Terahertz waves are a special kind of electromagnetic waves with a frequency range located between microwave and infrared waves [1]. In recent years, with the rapid development of terahertz technology and the improvement of industrial level, high-performance terahertz imagers have been successfully developed, which provide powerful technical support for close-range imaging. According to the different imaging modes, terahertz security imaging can be divided into active and passive. Active imaging systems obtain images by transmitting terahertz wave signals and detecting the echo signals scattered by the target. This imaging method is unaffected by temperature and surrounding radiation, and is highly adaptable to the environment. However, active imaging systems pose safety hazards to the human body and are relatively complex and costly. In contrast, passive imaging systems do not emit any rays or energy and utilise the difference in terahertz waves generated and reflected from the environment by the object and the body itself [2]. This imaging modality is simple in structure and has no radiation safety hazards, but the quality of the imaging may be greatly affected by environmental factors such as temperature and surrounding radiation.