Abstract

A sophisticated system for item detection and recognition is in high demand globally. The decision to focus on this research article was influenced, in part, by the increasing security concerns in Nigeria. Despite the implementation of various traditional methods to combat these issues, the threat continues to persist. Therefore, it is imperative to transition from outdated identification techniques to more contemporary ones. This study introduces a computer vision system using the Convolutional Neural Network (CNN) methodology with yoloV4 as the algorithm to offer a more efficient and effective approach to detecting human presence and verifying their identity. YOLO V4 is used to detect the human while CNN extracts the features of the human images from the recorded video. To assess performance, the system is compared against various models such as YOLO and YOLO V2. The proposed system demonstrated an identification accuracy of 72% on the MSMT17 datasets and a detection accuracy of 80.1% on the MS-COCO datasets, respectively.