Face Detection System for Public Transport Service Based on Scale-Invariant Feature Transform

Abstract

This paper proposed to reduce the complaints about the use of public transport. We applying the principles of digital image processing with the Eigen face detection and the Scale-Invariant Feature Transform (SIFT) matching technique. The system shown that the face of the person who interested will be mark and detect. After that it do an emotional analysis and show the emotional results immediately. We evaluated the effectiveness of our system based on the accuracy for detecting human faces. In the experimental, total testing 100 times with both of the straight and inclined faces. The results are as follows: A person’s face detection with a constant light is 90%, a 45-degree tilted face has a constant illumination of 79%, a 45-degree tilted face, a constant light of 55%, and a thin masked face section, it cannot be work.