Towards qualitative extraction of minutiae from fingerprints using some Image enhancement techniques

Abstract

ABSTRACT newlineExtracting minutiae features from poor quality fingerprints is one of the most newlinechallenging problems. It is extremely difficult for an automated system to newlineaccurately classify fingerprints which are poor in quality and to reliably locate the newlineminutiae. The present investigation deals with design of some image enhancement newlinetechniques to obtain reliable estimates of minutiae locations to facilitate qualitative newlineminutiae extraction. newline newlineThe work begins with the design of parameterized transformation functions that newlineuse local and global information of the image for the purpose of image newlineenhancement. Some new optimization techniques such as Modified Teaching newlineLearning Based Optimization (M-TLBO), Modified Harmony Search (M-HS), newlineSimple League Championship Algorithm (SLCA) have been explored for newlinecontrolling the parameters in the transformation functions in order to remove noise newlinefrom poor quality fingerprints. newline newlineThe work will then move to filtering with a view to further enhance the ridge newlinestructure. Modifications were made to some of the existing filtering techniques to newlineimprove the clarity of the ridge structure. A new methodology for effective newlineestimation of ridge orientation based on neural network and Ternarization process newlineis examined. The post processing step of removing false minutiae from the newlineAndhra University, Visakhapatnam newlineextracted minutiae based on some fuzzy type rules is also proposed and newlineimplemented. newline newlineA comparative study between the proposed techniques with many other available newlinemodels from the literature is done in order to determine the efficacy of the work. newlineThe relevant findings, improvements were outlined considering three evaluation newlinecriterions namely Fingerprint Image Quality, Robustness Index and Verification newlinePerformance. newline newline newline newline ABSTRACT newlineExtracting minutiae features from poor quality fingerprints is one of the most newlinechallenging problems. It is extremely difficult for an automated system to newlineaccurately classify fingerprints which are poor in quality and to reliably locate the