Novel Smile Feature Extraction Algorithm Using Improved Gabor for Mobile Phone Platform

Abstract

Due to the strengthening of the computer capabilities and storage capacity in mobile devices over the years, an incorporation of real-time smile recognition to mobile devices such as mobile phone and PDA is no longer unattainable. In this paper, we attempt to establish this possibility by presenting a novel face feature extraction algorithm in mobile devices for smile recognition. We extract global face features using Gabor filters, and then through a training process using the AdaBoost algorithm our system can efficiently search for key features in a large set of features. Finally we calculate the key points in the original image and the corresponding convolution masks. Experimental results show that the proposed algorithm significantly outperforms several popular smile recognition methods with a dramatic reduction in computational time.