Human Gait Indicators of Carrying a Concealed Firearm : A Skeletal Tracking and Data Mining Approach

Abstract

There has been an increase in crimes involving illegal firearms in the last couple of years. Previous studies have found that most illegal firearms are carried in a concealed manner. The detection therefore of persons carrying concealed firearms is critical in maintaining security especially in public places. Literature indicates that disruption in gait is a major indicator used by security personnel to detect persons carrying concealed firearms especially those tucked on the hip. However, the specific gait parameters that are indicative have not yet been quantitatively determined. The purpose of this study therefore is to analyze the gait of persons carrying a concealed firearm tucked on the right hip and to quantitatively determine the gait characteristics associated with carrying the firearm. A simulation of persons walking while carrying a concealed firearm and when unarmed was recorded using Kinect V2 depth camera. The depth camera provided 3D spatial skeletal joint position features of tracked joints for the armed and unarmed scenario. Paired t-tests were conducted to compare these features. Further, the results of the t-tests were related to the anatomical planes of Motion. Results showed that persons carrying a firearm demonstrated disrupted gait characterized by right arm abduction, left arm adduction, right leg adduction and extension. These findings extend existing gait indicators which can be employed by security personnel to identify persons carrying concealed firearms.