Performance Evaluation of Eye Detection System using WSAPG
Nwe Ni San, Nyein Aye · 2014
The precise face and eyes detection is crucial in many Human-Machine Interface system. The important issue is the reliable object detection method. In this paper we present the architecture of a face and eye detection system based on the Haar Cascade Classifiers and symmetry techniques. We use the feature based methods to detect the face and define the position of eyes. Our proposed system can accurately detect the center of both eyes in the rotated face image. In this system, the face region in an image is categorized by using color information in HSV space and is detected by using Haar Cascade Classifier. Then the face region in this image is extracted using color information in HSV space. After that we find the symmetric axis of the extracted face region using Weight of Symmetry Axis for Pixel Gradient (WSAPG) to detect the location of eyes. According to this axis we can determine the skewed angle of the extracted face region to rotate the face. After rotation process we find a horizontal eye line in the upper part of the face region by using the vertical symmetric axis and locate the center of eyes using the symmetry technique. By applying the proposed system to the set of 250 test images the 95% of the eyes were properly detected and precisely localized.