The Research Of The Fingerprint Image Quality Assessment Based On Gabor Filter

Automatic fingerprint identification technique has caused strong interest in biometrics field during the last decades. Due to the prevalence, convenience and reliability of authentication techniques and the complement and application of the fingerprint identification technique, the research of an automatic fingerprint identification system (AFIS) plays an important technical value and practical value. The automatic fingerprint identification system is a synthetic system that integrates photoelectric technique, image process, computer and network, database technique and pattern identification technique. It commonly includes fingerprint image collection, fingerprint image quality assessment, fingerprint image process, minutiae extraction, minutiae matching etc.The performance of the AFIS relies heavily on the quality of the input fingerprint image in application. The poor images will lower the performance of the AFIS. So fingerprint image assessment has become a crucial part and drawn a substantial amount of attention recently. This paper provides some researches on the fingerprint image quality assessment system.This article first introduces the basic principles of biometric technique and fingerprint identification technique, and then analyzes the existing method of fingerprint image quality assessment, the factors in the low fingerprint image quality, the cogitation of fingerprint image quality assessment, and the function of fingerprint image quality assessment module. We also introduce two typical methods of fingerprint image quality assessment, and the basic principles of the 2-D Gabor filter. Based on the characteristics of fingerprints and the need of fingerprint image quality assessment, we have design the 2-D Gabor filter with 8 directions. Finally, an efficient algorithm of fingerprint image quality assessment based on 2-D Gabor filter has been developed. In order to test our algorithm, a test system has been designed for fingerprint quality assessment research and a large number of experiments have been performed in it using FVC2004 standard fingerprint image database. All these results show that our algorithm achieve a good effect.