| CPC G08B 13/19613 (2013.01) [G06V 20/52 (2022.01); G06V 40/20 (2022.01); G08B 13/2402 (2013.01); G08B 13/246 (2013.01); H04N 7/18 (2013.01)] | 1 Claim |
|
1. A system for theft prevention and detection, the system comprising
a. a plurality of media capture devices capable of capturing multimedia from an individual activity at a work station and wherein the media capture devices are positioned on a premises boundary orthogonally or perpendicularly to the work station and wherein the media capture devices have a field of view of a target area comprising a counter and a cash register, and wherein the media capture devices have a field of view of head, arms, and hands of the individual interacting with the counter and cash register in the target area;
b. A plurality of sensors positioned on the premises boundary orthogonically or perpendicularly to the work station wherein the sensors are capable of capturing parameter data consisting of object weight, individual motion, object identification, individual velocity, light, and temperature of the target area;
c. A computer hardware architecture in wireless communication with the plurality of media capture devices and the plurality of sensors, the computer hardware architecture comprising a non-transitory computer readable storage medium having computer readable program code embodied therewith, the computer readable program code comprising computer readable program code for receiving, scanning, and processing data from the plurality of media capture devices and sensors, analyzing target object detection from data obtained by the plurality of media capture devices, and producing composite confidence estimates based on the sensor parameters to identify an instance of theft or misappropriation; and
d. operably configured code for a suspicious gesture recognition algorithm or for a hybrid sensor classification algorithm for detecting an instance of a theft or misappropriation wherein the sensor classification algorithm promotes processing of an updated target object count based on media captured during a video log, and wherein the suspicious gesture recognition algorithm promotes determination of a suspicious gesture or transaction by programmatically comparing context and an individual movement activities from a previous video log where the individual interacts with the counter and the cash register in the target area workstation;
wherein the computer readable program code for scanning initiates a scan process of a received plurality scan images from the plurality of media capture devices, synchronizes timing of the plurality of scan images from each of the plurality of media capture devices to produce a single composite scan image and produces a confidence estimate from the plurality of sensor data, updates the plurality of media capture device data with the confidence estimate of the plurality of sensor data and enhances the confidence estimate from multiple target object detection
wherein the composite confidence estimates are obtained through two rounds of processing of gesture data and object data from the plurality of media capture devices and sensors to indicate probability of a theft or misappropriation in real time by identifying individual gesture types of side pocketing, shirt pocketing, or hiding under the cash register.
|