Improved fingerprint identification/verification techniques using digital enhancement in the presence of noise and distortion.

摘要

Fingerprints are the most widely used physiological identifier in biometric recognition systems. Consequently, the domain of automated fingerprint matching is of utmost significance in guaranteeing reliable performance in automated fingerprint recognition systems. Successfully matching fingerprints is a difficult task due to the effects of noise and distortion that create large variability in different impressions of the same finger.; This work involves the application of image processing techniques to digitally enhance fingerprint images that have been subjected to noise and distortions. A correlation-based matching technique has been digitally implemented, using the concept of fringe-adjusted joint transform correlation.; This work shows that the Laplacian filter is the optimal image enhancement technique of the investigated spatial filtering methods. When used as inputs to the FJTC matching algorithm, the Laplacian-filtered images that have been subjected to noise and distortions yield significant correlation improvement in the verification and identification processes.