IMAGE ENHANCEMENT FOR FINGERPRINTS WITH MODIFIED TEACHING LEARNING BASED OPTIMIZATION AND NEW TRANSFORMATION FUNCTION

STEPHEN M.J.¹, REDDY P.V.G.D.^2
1Department of CSE, Wellfare Institute of Science Technology & Management, Visakhapatnam-530 027, AP, India.
²Department of CS&SE, Andhra University, Visakhapatnam-530 003, AP, India.

Received : 07-11-2013     Accepted : 05-12-2013     Published : 28-12-2013
Volume : 3     Issue : 1       Pages : 76 - 85
J Mach Learn Tech 3.1 (2013):76-85

Extracting minutiae features out of poor quality fingerprints is the most challenging problem. It becomes extremely difficult for the Automated Fingerprint Identification System to accurately locate the minutiae points in such fingerprint images. In the present work, an image enhancement technique is employed in order to obtain reliable estimates of minutiae locations prior to minutiae extraction. For the task of image enhancement a new parameterized transformation function is designed, which uses local and global information of the image. A new novel optimization technique, called Modified Teaching Learning Based Optimization (M-TLBO) is then implemented to control and change the parameters in the transformation function which is applied on the poor quality fingerprint images to remove noise. The best enhanced image is tried to achieve according to the objective criterion by optimizing the parameters used in the transformation function with the help of Teaching Learning Based Optimization (TLBO). So here Fingerprint image enhancement is considered as an optimization problem and TLBO is used to solve it. The work has been concluded with relevant findings based on improvements in Fingerprint Image Quality, Robustness Index and Verification Performance as three evaluation criterions. A comparative study between the proposed techniques with many other available models from the literature is done in order to establish the efficacy of the proposed enhancement techniques.