Published on October 2021 | Artificial Intelligence, Machine Learning, Video Analytics, Deep Learning, Neural Network, Internet of Things (IoT)

Anti-Theft system based on the Internet of Things (IoT) to monitor unusual movements.

When a homeowner or family member becomes aware of an ongoing theft, the suggested research approach seeks to provide a broadly applicable strategy for notifying them (unauthorized access to their premises). This was achieved by doing a thorough examination of current systems in order to identify research requirements. The previous systems' incapacity to differentiate between human and non-human items, as well as their ability to identify the thief only after the crime had occurred, were discovered to be problems. Wireless sensor networks (WSNs) are collaborating with the IoT and the Cognitive Internet of Thing- CIoT to make clever homes and their applications more widely available (CIoT). Researchers have developed a new smart home anti-theft system that can detect intruders even if their faces are partially or fully obscured by clothing or other materials such as leather, fiber, or plastic, according to the researchers. The suggested technique may identify a nocturnal intruder by using a CCTV. The overall aim of this project was to develop a low-cost, high-performance system that anybody could use to identify and report any kind of theft in real time, no matter how little. According to the creators, it would also provide real-time video data processing for home security systems. The study's findings indicate that the suggested method is feasible. When an intruder's face was partially or fully hidden, or when he was seen in the dark, the system's accuracy rose from 85.13 percent to 97.01 percent, 84.13 percent to 78.19 percent, and 66.51 percent, respectively. The accuracy increased from 85.13 percent to 64.13 percent, 56.70 percent, and 44.01 percent in these circumstances.