Preventing Child Kidnaping at Home Using CCTV that Utilizes Face Recognition with You Only Look Once (YOLO) Algorithm

Abstract

The rise of child abductions often occurs because of weak parental supervision and weak vigilance of strangers who intentionally or unintentionally approach children. Early warning when a child is approached by an unknown person can cancel the intention of the kidnapper. Indeed, CCTV can be used as a tool to thwart crime. Face Recognition is one of the best biometric recognition systems for identifying someone whose body or DNA has certain characteristics that distinguish one person from another. This research resulted in the right collaboration between CCTV which is equipped with the YOLOv8 algorithm which can detect a human object in real-time and also utilizes face recognition which can prevent child abduction by providing an early warning to kidnappers when approaching children. The results of the first test are the optimal combination using 3 cameras and the CNN method and the shooting time in the afternoon. This combination succeeded in providing an early warning in the form of a siren whose sound is issued with the help of a loudspeaker when a child is left alone and approached by an unknown person with the distance between the camera and the object to be detected as far as 1 meter to 3 meters which takes 15 seconds to 22 seconds to complete the detection process in an image. Then the results of further testing with 6 cameras using the LBPH method take between 17 seconds to 18 seconds to complete the detection process in an image, SVM takes 20 seconds to 42 seconds and CNN takes 17 seconds to 24 seconds.