LAUSR

Eyeglasses have a significant effect on the accuracy of face feature extraction and face recognition. In order to improve the performance of face feature extraction and face recognition, a method that achieves eyeglasses detection, location and frame discriminant, based on edge information is proposed in this paper. First, the horizontal nose bridge area is determined by the location of the mouth, and the vertical nose bridge area is determined by the edge information projection method. Next, the spectacle beam is searched in the nose bridge area to determine the existence of eyeglasses. Subsequently, the eyeglasses areas are located according to the bidirectional edge information projection method and the existence of frames is determined. Finally, the frame width of the eyeglasses is measured based on the location of the left and the right glasses. Experimental results show that detection rate of eyeglasses using the proposed method is higher than the traditional methods and it can make a distinction between rimmed and rimless eyeglasses accurately. In addition, the proposed method can locate the eyeglasses areas and measure the frame width effectively.