EDITH: AI for Detecting Accident Through CCTV Using MobileNetV2 with WhatsApp-WebJS as an Auto Reporting System

Leveraging technological advancements, the study focuses on addressing the critical issue of accidents that have lacked effective solutions. The major impact that will be caused by accidents is the increased death rate. The delay in handling accident victims is one of the causes of this increase. This research involves creating an accident detection model through the analysis of CCTV footage and integrating it into a user-friendly web application, which can give an alert of accidents to minimize the impact of accidents. The model is constructed based on MobileNetV2 architecture, with its performance enhanced using a more extensive dataset, achieving an impressive accuracy rate of 95 %. Python was utilized in developing the model, and subsequently exported to JSON format for integration with JavaScript. The choice of Node.js is due to its strong server-side capabilities and compatibility with TensorFlow.js, allowing us to use the pre-trained model on the client side. Additionally, this research integrates WhatsApp via the WhatsApp Web JavaScript library for easy messaging and media sharing on the website. The method successfully provides real-time alerts upon accident detection, automatically dispatching them as WhatsApp chat messages. Furthermore, a database stores the history of detected accidents, which can be accessed and displayed on the website. This approach has streamlined the monitoring of accidents detected by CCTV, making it effortless, automated, and accessible from anywhere.