Research and Simulation of an Improved Algorithm of Fingerprint Recognition

Abstract

Abstract. The application of biological recognition technology is more and more extensive [1]. Human body carries some unique biological characteristic such as fingerprint or sclera. On the base of the complexity of implementation, most of the enterprises choose fingerprint as the new Recognition way. Fingerprint Recognition technology has good safety and reliability. The traditional Recognition technology will be placed by fingerprint technology with the improvement of fingerprint Recognition. This paper introduces the basic process of the fingerprint Recognition technology includes the image pre-processing, the fingerprint feature extraction and feature matching. Image pre-processing includes image noise reduction, image cutting, and image binary. Feature extraction includes image thinning, minutiae extraction and fake feature point elimination. A look-up table algorithm based an "eight-neighborhood" is used in this step. The research sets an improved algorithm of three steps of matching: ridge line length matching, triangle length matching, and the point type matching. The simulation results indicate that this method has a high anti-noise performance and stability.