Fingerprint Image Enhancement based on Threshold Fast Discrete Curvelet Transform (FDCT) and Gabor Filters

Abstract

most important stages of Automatic Fingerprint Identification System (AFIS) are enhancement stage, features extraction stage, and matching stage. The main purpose of the enhancement stage is to increase the clarity of the fingerprint image, convert poor quality image to good quality image, and prepare the image for features extraction stage. Both of the two last stages (features extraction & matching stage) depend heavily on the enhancement stage, therefore this paper focus on the enhancement stage. Practically most of the input fingerprint images are corrupted by noise, body conditions, and environmental factors. Therefore it is necessary to use an effectively enhancement method. This paper present new efficient fingerprint image enhancement algorithm works by performing threshold on Fast Discrete Curvelet Transform (FDCT) domain and applying Gabor Filters. The proposed algorithm reduce the fingerprint image noise by using threshold fast discrete curvelet transform, then apply Gabor filters to enhance and increase the clarity of the image. The performance of proposed enhancement method is evaluated on the basis of Peak Signal to Noise Ratio (PSNR), Absolute Mean Brightness Error (AMBE) and the Computation Time (CT). Experimental results, and the comparison table between proposed enhancement algorithm against traditional methods at the end of this paper show that the proposed method is computationally efficient, with the same level of the enhancement performance. The proposed enhancement method is compared with other traditional fingerprint enhancement techniques. The performance is evaluated on the basis of Peak Signal to Noise Ratio (PSNR), Absolute Mean Brightness Error (AMBE) and the Computation Time (CT). Experimental results show that the proposed method is computationally efficient, with the same level of the enhancement performance.