Lips feature detection using camera and ARM 11

Abstract

This study aims to perform a simple lips feature recognition using lips angle detection. This feature detection starts from the search area of the face using skin segmentation continues to find the area of the lips by anthropometry human face in general. Then from the lips area to be searched its use lips angle detection algorithm to determine the coordinates of the corner of his mouth by the Lowest grayscale value combined with integral projections. Testing feature detection using 11 human faces, each doing four different expressions performed in real time. Four expressions are normal conditions, a thin smile, big smile and stare representing the possibilities of impending noise to ensure the accuracy of the lips features detection system. From the test results Obtained an average percentage of success of the system to normal conditions of 77.9%, a thin smile 91.68%, 95.97% and a wide smile.