Abstract
Optical fingerprint security verification is gaining popularity as it has the potential to perform correlation at speed of light. With advancement in optical security verification techniques, authentication process can be almost foolproof and reliable for financial transaction, banking, and etc. In law enforcement, when a fingerprint is obtained from a crime scene, it may be blurred and can be an unhealthy candidate for the correlation purposes. Therefore, the blurred fingerprint needs to be deblurred before it is used for the correlation process. There are a few different types of blur, such as linear motion blur, defocus blur induced by aberration of imaging system, and etc. Also, we may or may not know the blur function. In this paper, we proposed the Non-Singularity Inverse Filtering in Frequency/Power Domain for deblurring known motion induced blur in fingerprint. This filtering process will be incorporated with the Power Spectrum Subtraction Technique, Uniqueness Comparison Scheme, and the Separated Target and References Planes Method in the Joint Transform Correlator. The proposed hardware implementation is a hybrid electronic-optical correlator system. The performance of the proposed system would be verified with computer simulation for both cases: with and without additive random noise corruption.