Paper Title
Enhancing Taxi Driver Safety Through Security Eyes

Abstract
A taxi is a popular transport to carry passenger(s) to their desired destination(s). But nowadays taxi drivers are not secure due to the criminal passengers. Most of the time, they lose their lives by knife stabbing, shooting, or losing consciousness due to the chloroform. Previous researchers haven't worked on this issue. There are no safety measures or alert systems available for the drivers to protect them from potential criminal attacks by passenger(s) seated in the back seat. As a result, drivers are easily attacked by criminal passenger(s). The main aim of this paper is to reduce these heinous crimes and increase drivers’ security. That’s why the “Security eyes” system is introduced which consists of two camera modules, an alarm triggering system, a data transmitter, and internet connectivity. The system uses the YOLOv3 algorithm to identify suspicious activities (e.g., the use of gun, knife, or chloroform). One camera continuously observes passenger(s) activities. The second camera promptly activates and captures an image of the perpetrator(s), when any threat from the passenger(s) is detected by the first camera. Immediately, the system triggers an alarm to alert the driver and sends the taxi's location coordinates along with the captured image to the police facilitating easier apprehension by law enforcement. Based on the experimental results, the proposed system demonstrates a satisfactory performance rate with a high accuracy detection rate (around 97.82%). Keywords - Security Eyes, Object Detection Algorithm, Computer Vision, GPS