Fuzzy based Computing for Disabled using 3D Head Movement

For those with disabilities in the movement who have not yet had a fair chance to use the typical input devices of a personal computer, many authors have put forth replacements for the mouse during the past three decades. The overhead of employing head-mounted devices is decreased in camera-based systems by using the web camera as the mouse. The research problems and opportunities are tracking the user's facial expression of various users with various head poses through the camera and accurately converting into mouse cursor movement and click events. The usual GUI interactive features like menus and scroll bars that need horizontal or vertical movement are difficult to operate on the existing systems since they can only move the cursor in a slanting manner. Unintentional head movements by users frequently result in the current system losing track of facial feature tracking. The proposed system uses a standard web camera to capture the three-dimensional head rotation. To shift the mouse pointer vertically, the positions of the nasal bridge and nose tip are collected. The inner corners of the left and right eyes, as well as the tip of the nose, are also used to shift the mouse cursor horizontally. Unintentional head motions are disregarded to prevent the loss of face features, and by using fuzzy logic, the movement of the mouse pointer is only mapped with the intended head movements. The fuzzy control uses the head's rate, direction, and distance as inputs to determine how the mouse pointer will travel. By capturing the stable, purposeful, and sudden movements of the head, respectively, the fuzzy system classifies the movement of the head as weak, fair, or powerful. By removing the weak and strong head motions, the algorithm just maps the fair head movements with cursor movement on the screen. The proposed system has achieved the horizontal and vertical movement of the cursor and the results are significant when compared with the existing system. The system also successfully ignores the slight movement of the head captured by the web camera when the user remains the head stable, and the feature loss is completely avoided.