Research On Iris Recognition Algorithms

The increasing demand on enhanced security has led to an unprecedented interest in automated personal identification based on biometrics. Among the various biometric identification methods, iris recognition is widely regarded as the most reliable and one of the most active research topics in biometrics. Significant progress has been made since the concept of automatic iris recognition was first proposed in 1987, not only in research and algorithm development but also in commercial exploitation and practical applications. Most iris recognition systems consist of five basic modules: acquisition module, segmentation module, normalization module, encoding module and matching module. This thesis focuses on encoding module and matching module and covers the following work:1. Introducing an iris recognition algorithm based on improved SURF features in the thesis. According to the texture of iris. The original SURF feature extracting and matching methods were modified by finding the optical parameters in the algorithm. This modification enhanced the performance of the iris recognition system.2. Proposing a robust iris indexing algorithm using geometric hashing of SURF key points. The detected SURF key points were used to index iris database by applying geometric hashing scheme that is robust to similarity transformations as well as occlusion. It reduces the retrieval time and improves accuracy significantly for large scale database identification problem.3. A lot of experiments were implemented to find the best Gabor filters, which were used to extract the Gabor energy features of iris image. Proposed another iris indexing algorithm, which can also reduce the retrieve time and improve accuracy in large scale database identification.4. Implementing two proposed indexing algorithms in MATLAB. The algorithms were evaluated on CASIA Iris V4-Thousand database for the performance comparison.