An efficient Laplace-Beltrami Spectra based technique for online iris image compression and identification

Abstract

This research paper uses Laplace Beltrami Spectra based technique for online iris image compression for a person's eye. Normalized Eigen values (i.e. the spectrum) of iris image can be stored in the Smartcard memory and then it may be used for further identification. Hence, for verifying if two iris images are isometric or not, we need to compare the first ‘n’ Eigen values of the iris image spectra. If two iris images have the same Eigen values or same Riemannian Metrics values, it means both iris images are belonging to the same person. If two iris images have different Eigen values or different Riemannian Metrics values, it means both iris images are belonging to different persons. The experiments were conducted for 50 iris images of CASIA database. The experiments prove that the proposed technique is accurately identifying individuals on the basis of their iris images. The robustness testing was conducted by modifying few pixels in specific regions and few pixels in overall image. But, still the proposed method was able to identify individuals on the basis of their iris image patterns. The results of iris implementation reveal that the proposed method is an efficient and economically feasible.