Tsallis entropy, PCA and neural network in novel algorithm of iris classification

The apply of biometrics which refer to recognitize will become more widespread and trustworthy to develop, mature and provide identification with high degree of confidence in our technology. Classification of iris provides on important indexing technique in huge iris database. An accurate and coherent classification can be able greatly reduce time. The novel classification technique which is described in this paper is performed by Tssallis entropy. The proposed approach has been tested on 670 iris images which stored in CASIA database. In this paper we partition normalized image into 224 blocks and calculate Tsallis entropy of them, then we use PCA and neural network to classify images to six categories. Consequently based on purpose method, we can get encourage error rate classification (CCR) 97.31%.