Evaluation on a Method of Detecting Suspicious Objects with Transformer Model Using Passive and Active Imagers of W-band Radar

We are studying a method to improve the accuracy of detecting suspicious objects carried by walking persons using both passive and active imager images taken by the W-band radar system. In this paper, we propose a transformer model based on our previously proposed CNN-based model. According to the experiment, the proposed transformer model achieves better accuracy on suspicious object detection than our previously proposed CNN-based model.