Pre-processing And Segmentation Of Fingerprint Image Based On Image Quality Assessment

With the development of computer technology and information technology, biometric identification technology has been widely used. Biometric identification technology includes face recognition, fingerprint recognition, iris recognition technology, etc. Fingerprint recognition is the mainstream of biological recognition technology. Relative to the face recognition and iris recognition, fingerprint recognition has unique advantages, high practicability and reliability. In our country, fingerprint identification has been used for a long time. It has the advantages of convenient, safe, reliable, etc. And it has been widely used in areas such as financial security, e-commerce and digital encryption.Fingerprint identification technology includes four steps: fingerprint image collection, fingerprint image preprocessing, fingerprint image feature extraction and fingerprint image feature matching. This article introduced the fingerprint recognition technology background, the basic principle of fingerprint identification technology, and four steps of the fingerprint recognition technology.This article designed a method of fingerprint image preprocessing and segmentation based on the image quality evaluation. The purpose is lies in the quality assessment of low quality fingerprint image denoising, and according to the features of low quality image of the fingerprint image segmentation in order to get a clearly image segmentation. Before the fingerprint image denoising, using the fingerprint image quality assessment to select the low quality fingerprint images is the first step. In this paper, by understanding the fingerprint image acquisition process we knew the fingerprints of original image generally have a lot of noise. This is mainly caused by fingerprint image acquisition unit and the finger itself, such as the collector surface dirt, finger cuts, scars, fade skin, finger was dirty, dry, wet, etc. Therefore, an assessment of the fingerprint image quality is a necessary step. For fingerprint image is not clear for the follow-up work to bring trouble, this paper proposes a multi-target fusion quality evaluation method: the quality evaluation method based on fingerprint ridge valley line resolution. The method using the valley ridge line of the characteristic of the fingerprint image, and selecting effective fingerprint image area, multiple indicators for calculating grayscale average and variance as the fingerprint image quality evaluation factors, and through the grades of each evaluation factor, the fingerprint image quality score is calculated by weighting.This paper using the wavelet threshold method to deal with the noise of fingerprint image. Wavelet denoising is a signal filtering problem, when used in image denoising can restore and preserve the original image, so this article consider adopting wavelet denoising. Through quality evaluation and wavelet denoising, after entered into the phase of the fingerprint image segmentation. The object of this study is low quality fingerprint image and the purpose is of low quality fingerprint image segmentation effectively. Therefore, according to the characteristics of the low quality fingerprint images, this paper presents a feature analysis of fingerprint segmentation method. This method takes advantage of the fingerprint image characteristic analysis of the fingerprint image, respectively, from the fingerprint image’s gray characteristic and direction for segmenting the fingerprint image with multiple features fusion, overcame the disadvantages of two kinds of methods. And it can accurately separate the background area from the fingerprint image, improve the accuracy of the segmentation.This paper design the experimental process on Windows platforms with Matlab7.0programming implementation, the experimental results show that the set of methods to deal with the fingerprint image segmentation effect is good, can effectively for fingerprint image segmentation.