Classification of Threats Via a Multi-sensor Security Portal

Abstract

Classification of threats has always been a major security decision issue when a person walks through a security portal. To classify threat by combining data from multiple sensors like Millimeter Wave (MMW) imagery and Multi-zone Metal Detector is proposed and studied. Decisions made according to data from multiple sensors are analyzed. We present results for processing the fused sensor data and study the performance of several learning algorithms on this task. Usually the decision is being made based on single types of sensors, e.g., pulse induction metal detectors. We wish to identify the threats which are not determined by metal detectors alone.