A novel cancelable iris recognition system based on feature learning techniques.

We present a novel cancelable iris recognition system.Based on different feature learning techniques followed by Spatial Pyramid Mapping technique is employed and investigated for feature computation of iris pattern.For the proposed system the existing BioHashing technique is modified that use user specific and user independent tokens.Six benchmark iris databases namely: MMU1, UPOL, CASIA-v3-Interval, IITD, UBIRIS.v1 and CASIA-syn are used to test the performance of the proposed system. A novel cancelable iris recognition system is proposed in this paper. Based on the performance of various feature learning techniques such as (i) Bag-of-Words, (ii) Sparse Representation Coding and (iii) Locality-constrained Linear Coding we choose the second one followed by Spatial Pyramid Mapping technique for feature computation from iris pattern. To build the proposed system the existing BioHashing technique is modified using two different tokens: one is user specific and the other user independent. To test the performance of the proposed system we have tried it on six benchmark iris databases namely: MMU1, UPOL, CASIA-Interval-v3, IITD, UBIRIS.v1 and CASIA-syn. The experimental results are demonstrated for each database and are compared with that of the state-of-the-art methods with respect to these databases. The results show the robustness and effectiveness of the proposed approach.