A Novel Approach towards Iris Segmentation and Authentication using Local Chan-Vese Method

Iris segmentation has been an especially interesting research area from the last decade due to the increased security conditions for the sophisticated personal identification ideas based on biometrics. The rich distinctive and stable textural information of the iris models make iris a biometric modality for identifying each person correctly and reliably. Most recent iris segmentation techniques show the high segmentation accuracies in cooperative environments. However, the iris image segmentation remains a difficult topic. In this frame work, we proposed an innovative model as an improvement of Chan-Vese technique by incorporating B spline approach to perform iris segmentation. Proposed scheme has added enhanced segmentation for non-ideal iris images in visible light. The GLCM (Gray Level Co-occurrence Matrix) and LBP (Local Binary Pattern) are employed for feature extraction. This scheme is able to perform all the associated treating in 1-dimension as the B-spline task is divisible and is built as the result of n-1), 1-D, B-splines. This presents superior control compared to other methods. Experimental results displays that the proposed iris segmentation technique considerably minimizes the required time to segment the iris without affecting the segmentation precision. The main benefits of this algorithm are: First, it can deal with the accurate recognition of smooth objects. Second one is, it can powerfully handle the noisy images. Therefore, thereal boundaries are conserved and correctly distinguished Additionally the comparison outcomes with related iris segmentation methods show the superiority of the proposed work in terms of segmentation accuracy and recognition performance. The NICE I iris image database is used to compute the performance of the proposed technique.