A High-Precision Target Detection Algorithm Based on a Small Amount of Dataset

Millimeter-wave body scanners have emerged as a key technology in modern security inspection systems. However, the limited availability of datasets for training models in this field makes it challenging to achieve high-precision detection. In this paper, we propose an optimized approach for detecting dangerous objects using a limited dataset. By leveraging YOLOv5 and YOLOv8 models, the proposed method can achieve an accuracy of 75.1% even with a limited dataset, ensuring practical applicability for real-world security scenarios. The results demonstrate the effectiveness of the proposed approach in maintaining high detection performance despite data limitations.