Font Size: a A A

Dsp-based Fingerprint Identification System

Posted on:2011-05-16Degree:MasterType:Thesis
Country:ChinaCandidate:Z Q LiaoFull Text:PDF
GTID:2208360308471815Subject:Circuits and Systems
Abstract/Summary:PDF Full Text Request
Fingerprint identification is an authentication technology based on biometrics recognition. Due to the uniqueness and invariability of fingerprints, the automatic identification based on fingerprints is becoming an alternative to the traditional methods of identification. With the extensive application of DSP in image processing, the fingerprint identification system that uses fingerprint sensor and DSP as the core will popular in future.According to the special applications of fingerprint identification in the DSP system, automatic fingerprint identification algorithms are studied, and the transplantation of pre-processing algorithm of fingerprint image from PC to the DSP system has been completed. The paper's major works are as follows:1. Based on the local variance information of fingerprint image, a new algorithm for calculating fingerprint directional image is proposed in this paper. The proposed algorithm is suitable for embedded systems with little calculation and easy to implement. According to the fingerprint directional image, a group of directional filters is constructed to enhance the image. At last, the enhanced image is binary by fingerprint directional image.2. Fingerprint matching is an important step of AFIS. For the problem of locating reference point, a new fingerprint matching algorithm is proposed based on the feature points in the central region of the fingerprint image. Firstly, the matching points are obtained as reference points by early match, and then, the global coordinate of all of the feature points is adjusted, at last, multi-discriminant is integrated to complete fingerprint matching.3. The target of the paper is to make a DSP automatic fingerprint identification system. Firstly, the automatic fingerprint identification algorithms are studied in the VC and MATLAB environment, and then, transplanted from PC to the DSP system. The solutions used in the paper is describes. The actual test results indicated that the algorithms are correct and feasible on the DSP.
Keywords/Search Tags:Fingerprint, Orientation Filter, Matching, DSP
PDF Full Text Request
Related items