A New Method for Tenprint Image Quality Evaluation

A new method is proposed for tenprint image quality evaluation in this paper. The method can be employed to evaluate the quality of tenprint images which are captured by a live-scan scanner. First, the method evaluates whether there is gray inversion in the tenprint image by comparing the average gray values of the sub-images of four corners of tenprint image with the maximum gray value of tenprint image. Second, the tenprint image is projected in the horizontal direction, which can evaluate whether there is a large offset between the core of fingerprint friction ridge area and the center of the tenprint image. Third, a new method based on two-dimensional weighted block DCT (Discrete Cosine Transform) is proposed to evaluate the sharpness of tenprint images. The sharpness evaluation function is defined in this paper. Big weights are given to high-frequency components and small weights are given to low-frequency components in the sharpness evaluation function. Fourth, the sharpness evaluation function is used to evaluate whether there are blurred areas in the core area and the delta areas of the tenprint image. Eventually, the sharpness evaluation function is used to evaluate whether there are blurred areas or incomplete areas outside the core area and the delta areas. The method in this paper is tested on some common tenprint images. Experimental results show the method in this paper is effective and feasible.