PERFORMANCE ANALYSIS BETWEEN KEYPOINT OF SURF AND SKIN COLOUR YCBCR BASED TECHNIQUE FOR FACE DETECTION AMONG FINAL YEAR UTEM MALE STUDENT

Abstract

This project presents an analysis of development out-of-plane of face detection using Speeded-Up Robust Feature (SURF) and skin colour of YCbCr colour-based technique. The technique of SURF method and skin colour of YCbCr was explored in order to compare the performances in terms of time response. The significant difference can be seen with an extraction of feature point followed by matching it using SURF technique whereas skin colour of YCbCr colour extract the skin region. It is discovered that the skin colour of the respondent does not give any impact on the result. The outcome is presented using MATLAB 2013a software. To determine the response of both technique in detecting the face area, out-of-plane captured images is varied and chosen randomly from 0°, 45° and 90°. The outcome shows that SURF technique can detect the SURF feature point in different angles, but the matching point cannot discover if the images in 45° and 90°. In contrast, the skin colour of YCbCr can spot the present of face despite its angle. Through the project analysis, the respondents’ tone of skin colour does not affect the result of both techniques. Overall, SURF technique gives an impact in face detection if the angle of the images is being varied. Different angles are applied in this technique in order to vary the result of out-of-plane. The number of key feature for image 2 is the highest due to variation of angles from 90°, 45° and 0° in corresponding.