<![CDATA[The more sophisticated technology makes all electronic devices can be made through joysticks, remote controls, and so on. This is a challenge for stroke sufferers because it limits them to move the member must make it able to rotate the device. Then the menu selection system needs to be made on the display that is able to make a sound to facilitate sufferers in communicating. Stroke sufferers have limitations to move members but can still move the two balls. This research was made to overcome their limitations by using the K-Nearest Neighbor classification method to classify the value of features resulting from ball motion detection using digital image processing with the face landmark method to convert eye areas and using the LBP method to extract features in the eye area. This system produces an accuracy of 100% and in dim lighting produces an accuracy of 60%, 20%, 100%, and 100% for moving the eyeball forward, right, left, and up. 100% accuracy results. The results of computational time are 399.7 ms, 398.4 ms, 398.4 ms, 396.8 ms for moving the eyeball forward, right, left, and up.]]>
Copyrights © 2020
