An Improvement of Iris Pattern Identification Using Radon Transform
Pattraporn Ariyapreechakul, Nongluk Covavisaruch · ECTI Transactions on Computer and Information Technology (ECTI-CIT) · 1970
This research proposes an improvement of iris personal identification system using Radon transform which is used to extract feature stored as templates in a database. In our research, the system has been tested to find the best system performance with 3 distance functions: regional correlation, Euclidean distance and absolute distance. The test images are from CASIA iris database. Our experimental results reveal that the improved approach greatly increases the system performance for all distance functions. Among the 3 distance functions, the absolute distance gives the best results which reduce EER from 18% (previous approach) to 3.695%.