In order to evaluate our system performance, we have used the public iris database CASIA [10], which is a most used benchmark in the iris recognition research.In the evaluation process, three images from each person have been taken as a reference and four as test. In order to compare the LBP method and the proposed NBP method, twenty persons taken randomly from the database are used in the experimental process. Thus, 60 images are used as references and 80 as a test images. Each test image is considered as query. The LBP histogram is extracted and the mean variation of the NBP image is extracted. After that, the hammingdistance between the query’s feature and the features extracted from all reference images of the database is calculated. The obtained hamming distances are sorted from the most similar to the dissimilar comparing to the query. A majority vote of the top three is considered and the query iris is classified following the majority. An experimental example is illustrated in Fig 6.