Eyes Detection on RGB Image for Controlling Side Mirror

Abstract

Eye detection is a process where the position of the eyes is extracted from the face images. This is one of the methods that are used nowadays for many applications such as for security purpose and drowsiness detection. This project proposed the use of eyes detection for controlling side mirror of motor vehicles. This eyes detection method aims to make driving easier without need to manually adjust the side mirrors. The system will find the midpoint coordinate of eyes detection on RGB (color) image and the input signal from y-coordinate will transfer to controller in order to rotate the angle of side mirror on vehicle. Viola Jones detection and circular Hough transform are used in image processing to detect driver eyes on RGB image. The coordinate of midpoint eyes is carried out from the experiment and been testing using controller that has been provided in order to find the angle of adjusting the side mirrors. The graphical user interfaces (GUI) was designed to display the output image and the findings of midpoint. The detection was detected from the part of position eyes on the face image. The position of eyes was cropped and the coordinate of midpoint was successfully comes out from the circle iris detection. The results using function of Circular Hough Transform on the images that have been selected, shows 80% have the same sensitivity value.