Research On Image Quality Assessment And Preprocessing Of Fingerprint Identification System

With the in-depth development of information technology, digital and network, network security and information systems security are very concerned about. Identity authentication as the first gate of a network security and information systems security, the security strength of it direct impact on the level of security of the entire security system. The traditional authentication methods that based on Symbol or logo can not meet today's demands on security because of its inherent defects. Biometric technology which has unique, stability, ease of use and reliability, is considered future direction of network security and identification. In all biometric technologies,fingerprint recognition technology becomes the best authentication technology because of its unique advantages.At present, many scholars, research institutions and companies have in-depth research on the fingerprint recognition technology. However, due to the lack of systematic and holistic, it is difficult to improve the performance of fingerprint recognition system, which limiting the wide application of fingerprint identification system. Fingerprint image quality assessment and preprocessing play an important role in fingerprint recognition technology. The paper has in-depth analysis on these two fingerprint recognition technologies, and researches from the following two aspects:on one hand how to select evaluation parameters to evaluate the input fingerprint image quality and thus strictly control the quality of the original system input; on the other hand how to organize, optimize and innovate out of an effective, standard fingerprint preprocessing algorithm to improve the fingerprint preprocessing speed and accuracy.Therefore, the paper revolves the topic of the fingerprint image quality assessment and preprocessing. Through analyzing and sorting out the previous proposed algorithms, the paper has in-depth research on image quality evaluation of parameter and process, the fingerprint image pre-processing. This paper's contents and innovations are as follows:(1)At present, image quality evaluation system lacks a generally accepted standard, it is still at the exploratory stage. This paper proposes a set of fingerprint image quality evaluation quotas, including the area quality score, location offset quality score, dryness quality score, humidity quality score and fuzzy direction quality score. Based on these this paper designs a fingerprint image quality evaluation process, which is the combination of steps excluding and comprehensive evaluation.(2) In order to solve the lack of systematic and holistic problems in fingerprint image preprocessing, this paper designs an effective, standard fingerprint image pre-processing algorithm flow. Including multistage segmentation system based on segmentation of adaptive threshold; the one-way multistage median filter; image equalization; the oriented filtering enhancement based on query table; a smooth dynamic local threshold binarization using the orientation information based on block;a mixed parallel-serial thinning algorithm based on improved OPTA(One-Pass Thinning Algorithm).Many of these algorithms are the forefront of international research, and many algorithms are integrated into my own point of view.(3)These algorithms are modularly designed. In order to improve the effectiveness of those algorithms,the flows of the algorithms have been optimized design.