Abstract
This thesis consists of two parts: eye wink control interface and iris recognition. In the first part, we developed a user interface based on eye wink control scheme. We apply the support vector machine and template matching algorithm to detect and track eye winks. After that, the dynamic programming is used to estimate the input commands. Thus, users can control the computer-based device according to their varying duration of eye winks.In the second part, we propose an iris recognition method by using the matching pursuit algorithm to extract the most significant features of iris. The feature extraction includes two parts: iris location and feature extraction. We apply the matching pursuit algorithm to extract the most significant features, and eliminate the unnecessary information in order to reduce the dimension of iris signal. The identification of irises is based on the similarity between the corresponding feature vectors. The experimental results show the performance and efficiency of our proposed framework.