Abstract:Intelligent recognition techniques, plays a significant role in multidisciplinary problem solving. In particular, various problems like character recognition, iris recognition, fingerprint recognition, signature recognition, and many others can be intelligently solved and automated.The objective of this thesis is to design and build iris recognition system which can be able to recognize the individuals by using artificial neural network. The iris recognition system consists of an automatic segmentation system that is based on the Daugman method. The extracted iris region was then normalized into a rectangular block with constant dimensions to account for imaging inconsistencies. The enhancement was made on the normalized image and then a technique was proposed for noise that includes eyelashes, lighting reflections and specular reflections. Next a stretch technique was used to stretch the histogram of the enhanced image. The next step is to quantize the iris to 16 levels. Then, the texture approach (co-occurrence matrix method) was used to extract the features of the iris to create iris template and store the results in a database. Finally, the probabilistic neural network (PNN) was employed for classification of iris templates. The designed system is implemented using Matlab V7.6 (R2008a) and tested on a set of iris images gathered from 36 different persons, 5 samples for each person. The experimental results show that the proposed system has high accuracy (100%).