Title :
An Effective Approach to Fingerprint Segmentation using Fisher Basis
Author :
Afsar, F.A. ; Arif, M. ; Hussain, M.
Author_Institution :
DCIS, PIEAS, Islamabad
Abstract :
Fingerprint image segmentation is an integral part of an automatic fingerprint recognition system. In this paper an effective algorithm for fingerprint image segmentation is presented. This technique is based on the fusion of multiple features for segmentation that are projected onto a one dimensional feature space using Fisher discriminant analysis. The classification of the fingerprint regions as foreground or background is carried out by the use of learning vector quantization (LVQ) neural networks. The proposed method exhibits low segmentation error (1.8%) and its high performance makes it an excellent candidate for use in online fingerprint verification systems
Keywords :
feature extraction; fingerprint identification; image segmentation; learning (artificial intelligence); neural nets; vector quantisation; Fisher discriminant analysis; automatic fingerprint recognition system; fingerprint image segmentation; learning vector quantization neural networks; one dimensional feature space; online fingerprint verification systems; Background noise; Degradation; Feature extraction; Fingerprint recognition; Fingers; Gabor filters; Image matching; Image segmentation; Image sensors; Neural networks; Fingerprint Identification/Verification; Fisher Discriminant Analysis; LVQ Neural Networks; Segmentation;
Conference_Titel :
9th International Multitopic Conference, IEEE INMIC 2005
Conference_Location :
Karachi
Print_ISBN :
0-7803-9429-1
Electronic_ISBN :
0-7803-9430-5
DOI :
10.1109/INMIC.2005.334432
