With the rapid development of information technology and multimedia technology, the information that people have, like image or video, is more and more. Image recognition has already been widely researched and applied in recently years, and face recognition is playing a very important role in it. It has a good prospect in identity recognition. But due to the particular structure of face data, there still have many problems in this field. We must do more deep study in order to reach the ideal recognition rate. Therefore, a novel algorithm which introduces Two-Dimensional Principal Component Analysis into Independent Component Analysis is proposed in this thesis. It is called 2DPCA based ICA face recognition.First, the technologies about Principal Component Analysis, Linear Discriminant Analysis and Two-Dimensional Principal Component Analysis are introduced. Secondly, the Independent Component Analysis and its application are introduced. The merits of this algorithm are as follows: the method introduced 2DPCA into ICA improves the learning speed. In the small sample case, we can get higher dimensions than classical PCA. We can also improve the recognition rate using this alagrithm. At last, the algorithm is analyzed theoretically, and experiments are carried on to test the effects of recognition algorithm based on given data. |