کنفرانس مهندسی برق ایران

صفحه اصلی / سی و چهارمین کنفرانس بین المللی مهندسی برق

A Robust Algorithm for Human Presence Detection in Millimeter Wave Imaging Systems

نویسندگان :

Arman Miraghasi Khani¹ Mohammad Hossein Gharib² Mahdi Shabany³ Zahra Kavehvash⁴

1- دانشگاه صنعتی امیرکبیر 2- دانشگاه صنعتی شریف 3- دانشگاه صنعتی شریف 4- دانشگاه صنعتی شریف

کلمات کلیدی :

artificial intelligence،human presence detection،millimeter wave body scanner،object detection

چکیده :

The systematic screening of individuals in densely populated and security sensitive environments is of paramount importance. Millimeter wave (MMW) body scanners represent one of the advanced technologies currently employed for this purpose. A key advantage of these systems is their ability to automatically detect objects concealed beneath clothing through the use of deep neural networks. In cylindrical millimeter wave (MMW) imaging configurations, antennas are mechanically rotated around the subject’s body by an electric motor. However, minor mechanical instabilities prevent the complete elimination of environmental clutter. As a result, when the system operates without a subject present, residual noise may be introduced into the images, potentially leading to false alarms in object-detection neural networks during empty scans. This study presents a robust algorithm designed to determine whether a person is present in MMW images. When human presence is confirmed, the corresponding images are processed by deep neural networks for automated detection of concealed items. Conversely, if no subject is detected, the system issues an alert to the security operator, indicating that the scan was conducted without an individual inside the device. Experimental evaluations indicate that the proposed algorithm delivers excellent performance, achieving 100% recall, precision, and F1-score on the test dataset. In addition, the algorithm requires minimal testing time and operates efficiently in real time.