Research Of The Fingerprint Recognition System Base On Image Content And Its DSP Implementation

Automatic fingerprint identification system (AFIS) has integrated image processing, pattern recognition and database technique. Tracing back to 1960's when it' s first introduced, AFIS has been successful and still the hotspot in pattern recognition domain today after 40 years' development.First, this paper introduces the advantages and disadvantages of three types technology of fingerprint scanning, and use the fingerprint sensor MBF200 for capture image.Then, aim at self-designed fingerprint scanning device, a set of algorithms for fingerprint recognition base on image content is presented, which include fingerprint image calibration, preprocessing, and feature extraction. By center locating, achieved image calibration; and the preprocessing method include segment, ridgeline optimization, thin and so on. This preprocessing method can protect ridgeline completeness and smooth edge; Used Mahalanobis distance for feature extraction. Implementing the algorithm using of Matlab6.5, and this set of algorithms achieve good recognition rate. So this set of algorithms for fingerprint recognition base on image content simplifies the process of feature extraction. Provide a feasible method for the fast fingerprint recognition.This paper achieved codes transplantation and optimization on DSP, and do something about codes optimization. The C6713 DSK which belongs to TI is chosen. Used CCS, successfully build a AFIS.