Iris feature extraction and classification based on serial images

作者：Tian Qi Chuan^*; Liu Zheng Guang; Li Lin Sheng

来源：Journal of Tianjin University, 2007, 40(12): 1441-1446.

• Feature projection
• Feature subspace
• Iris classification
• Serial images

In iris recognition system based on single image, there are many features changed by interferences and validity of these features cannot be judged from single image. So it is difficult for recognition system to make correct iris classification. To solve the problem, an algorithm of iris feature extraction and classification based on serial images was proposed in the paper. Firstly, iris binary ordinal features were extracted from several images. Then, iris features'; positions were aligned by finding maximum similarity of iris features and the iris feature subspace was achieved from serial iris features'; projections. Finally, iris identification could be achieved by calculating similarity between the feature subspace and those features in database. Compared with the method based on single image, the iris classification performance of the proposed method is obviously improved.

• 出版日期2007
• 单位天津大学