Theft Detection using Deep Learning
Sairaj Shirole · Research Square · 2023
Abstract In earlier times, crime detection relied solely onhuman observation, lacking efficient methods for detection. Theadvent of CCTV cameras marked a significant advancement incrime detection, but the manual review of video footage byhumans proved to be a time-consuming process. In presentworld, Artificial Intelligence (AI) and Machine Learning (ML)have made significant strides, the need for intelligent systems toautomate crime detection in CCTV surveillance has becomeparamount. Such systems can not only detect crimes but alsoclassify them and provide alerts to nearby police stations andambulances, thereby contributing to the reduction of crimerates in any given country. Object detection and tracking incomputer vision have gained widespread attention due to theirdiverse applications, including surveillance and securitysystems. Researchers have diligently worked to improve theaccuracy and efficiency of these processes. Our system aims toenhance security measures and facilitate swift responses topotential threats by employing real-time object detection on livevideo feeds. Furthermore, this system can be further optimizedthrough the integration of specialized hardware, ensuring evenmore robust and efficient crime detection capabilities.