For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
A Fast Automatic Fingerprint Identification Method Based on a Weighted-Mean of Binary Image
Yu HE Ryuji KOHNO Hideki IMAI
IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences
Publication Date: 1993/09/25
Print ISSN: 0916-8508
Type of Manuscript: Special Section PAPER (Special Section on Information Theory and Its Applications)
automatic fingerprint identification, pattern recognition, weighted-mean, optimization, personal identification,
Full Text: PDF(1.1MB)>>
This paper first proposes a fast fingerprint identification method based on a weighted-mean of binary image, and further investigates optimization of the weights. The proposed method uses less computer memory than the conventional pattern matching method, and takes less computation time than both the feature extraction method and the pattern matching method. It is particularly effective on the fingerprints with a small angle of inclination. In order to improve the identification precision of the proposed basic method, three schemes of modifying the proposed basic method are also proposed. The performance of the proposed basic method and its modified schemes is evaluated by theoretical analysis and computer experiment using the fingerprint images recorded from a fingerprint read-in device. The numerical results showed that the proposed method using the modified schemes can improve both the true acceptance rate and the false rejection rate with less memory and complexity in comparison with the conventional pattern matching method and the feature extraction method.