Fingerprint Image Minutiae Extraction

With the development of our society, the traditional security system, based mainly on token or password, becomes more and more fragile and can no longer meet the demands of the modern security system. With its uniqueness and invariability, the fingerprint has been used as a main evidence to identify individuals. It is superior to other biological features with its special virtues, among which the most important are its practicability and feasibility. Considered the main method of the biological feature recognition, the fingerprint technology is embracing a bright future.This paper focuses on the study on the fingerprint image's preprocessing and minutia extraction approach under the automated fingerprint identification systems (AFIS), which is based on minutia features. This paper mainly discusses the following three research aspects:(1) Computing the orientation image of a fingerprint image. After an intensive analysis and comparison of several methods of computing the orientation image, we select the "based on image pyramid" method in our research. The proposed technique is more robust against image noise than others.(2) Segmentation of a fingerprint image. This paper discusses the advantages and disadvantages of two segmentation algorithms, which are respectively based on variance of gray and direction image. When the two algorithms are integrated for the segmentation of the fingerprint image, better result is achieved.(3) Enhancement of a fingerprint image. A fingerprint image enhancement method based on Gabor filter is presented and realized. In our algorithm, after Fourier transforming transforms the fingerprint enhancement from spatial domain to frequency domain, the fingerprint images are enhanced by directional filter in frequency domain.All algorithms are implemented on PC under Matlab. The results demonstrate that the processing effects of fingerprint images can be greatly enhanced.