Paper Title
Facial Landmark-Based Cursor Automation And Drawing Interface

Abstract
It is a project concerning a computer control system which is capable of being operated in such a manner that there is no physical contact with it. It tracks your movements using a web camera like the eye blinking and mouth opening to use a mouse cursor and drawing desktop. The technology reacts to your gestures and uses machine learning and computer vision to understand you. This can help the motor challenged individuals easily use the computers. It is also a channel of interaction of human beings with computers. The program is based on the dlibs facial landmark detector. The tools help it to identify the parts of your face like your mouth and nose, mouth and nose. Its working is as below, It takes two ratios the Eye Aspect Ratio (EAR) and the Mouth Aspect Ratio (MAR) with time. These ratios assist it on the detection of eye blinks and opening of the mouth. The first one is the calibration stage. This stage involves the system identifying your motions and setting specific boundaries to identify your gestures. Then it reads your movements like the movements of a mouse. Right and left mouse clicks are activated with blinking left and right respectively. Mouth openings are observed to alternate between scrolling and drawing. The movements of the head control the cursor on the screen. It is also within a system of drawing. The prototype has proven that the gesture-based control is good in this mode whereby one can create a drawing on a canvas just by moving his or her head. It consists of landmark tracking, adaptive thresholding and mouse emulating event. This enables it to react to the mouse clicks depending on the blinks of head position This was initially proven to be correct not only in lighting but also in response to moderate head movements. This project shows that vision-based assistive technology may be used to improve accessibility, creativity, and the user experience. Keywords - Computer Vision, Facial Gesture Recognition, Human-Computer Interaction, Mouse Controller