Fingerprint Image Enhancement Algorithm Based On Steerable Filters And System Realization In FPGA

Currently, fingerprint verification has become the most widely used in Biometricfield. The fingerprint image enhancement technology is the core of the fingerprintimage pre-processing, which plays an important part in recognition technology. Untilnow, there are many fingerprint image enhancement algorithms with quite goodperformance; however, these algorithms are not suitable for applying in the increasingminiaturized mobile fingerprint recognition devices. Based on the above backgrounds, ahigh-performance fingerprint image enhancement algorithm which can be used inembedded system is proposed in this paper. It mainly includes the following two parts:1. Fingerprint image enhancement algorithm based on Steerable filters has beenproposed, which includes image contrast enhancement, fingerprint orientation imageestimation, image smooth, frequency image estimation, orientation image estimationbased on Steerable filters and Steerable filtering, etc. Its innovation is to improveSteerable filters and to apply it in the fingerprint orientation field extraction andfingerprint image enhancement. In addition, the directional separation characteristics ofSteerable filters is suitable for parallel processing in embedded systems, such as FPGA.2. A fingerprint image enhancement system based on FPGA has been implemented,including the fingerprint image acquisition, image enhancement and image display. Thesystem is mainly designed in Verilog HDL. Two-dimensional look-up table has beenused to achieve trigonometric functions; it can help to avoid floating-point operationsand large capacity memory required to store all the possible results in usingone-dimensional look-up table method.