Study On The Analyzable Method Of Fingerprint Morphological Features

Fingerprint morphological features are important basis for studying fingerprint identification system.The fingerprint feature extraction algorithm based on morphology can only extract a single feature,and to obtain a variety of features requires a combination of different algorithms.Can we simultaneously extract a variety of fingerprint features by a single algorithm,which can not only reduce the computational complexity,but also save the system time cost?In this paper,an analyzable method of fingerprint morphological features is proposed to solve the problem of decreasing accuracy and high time cost of traditional algorithm to extract a variety of fingerprint features,so as to achieve the purpose of simultaneously extracting a variety of fingerprint features with one algorithm.Firstly,the paper studies the analyzability of fingerprint morphological features,establishes the connection between fingerprint image and twodimensional signal,and regards the fingerprint image as a two-dimensional AMFM signal according to the texture structure characteristics of fingerprint.Secondly,the meanings of each component in the fingerprint signal are analyzed,the highfrequency noise and gray compensation value in the fingerprint signal are eliminated by gaussian low-pass filtering and mean filtering,and the front and back background of fingerprint image is separated by image segmentation to obtain the fingerprint signal containing only amplitude and phase information.Thirdly,the two-dimensional analytical signal of fingerprint signal is constructed.According to the properties of analytical signal,the amplitude and phase of fingerprint signal are separated,and the orientation field and phase field of fingerprint signal are obtained.Finally,according to the relationship between the components of two-dimensional AM-FM signal,other characteristics of fingerprint are obtained through the orientation field and phase field.The experimental results show that the algorithm proposed in this paper can simultaneously obtain a variety of fingerprint features by analyzing fingerprint signals and the relationship between the components of two-dimensional AM-FM signal,which not only improves the accuracy of fingerprint features,but also saves the system time cost and improves the performance of the fingerprint recognition system.