Thumbprint preprocessing using altera cyclone DE2 board

Abstract

Thumbprint has been by far the most reliable and acceptable form of biometric identification system. This project aims to implement the thumbprint preprocessing procedures from the software front with the utilization of Altera Cyclone DE2 board to improve raw thumbprint images up to a standard necessary to facilitate subsequent processing processes for thumbprint recognition. In order to realize this, the configurable FPGA hardware of the board has to be ready to enable the operation of the Nios II processor. The preprocessing procedures have to be defined in a description (ANSI-C++) that is understandable by the board. This project looks out for an implementation option that satisfies the requirements on flexibility and portability while trying to incorporate as much as possible the needs from multiple perspectives, namely computational cost, algorithm speed, and storage requirement. The pre-processing procedure generally comprises five main stages, namely filtering, contrast enhancement, adaptive segmentation, noise elimination and thinning. The project has proven that Altera Cyclone DE2 board is a feasible implementation option for thumbprint image preprocessing. Comparable image quality can be achieved with the thumbprint preprocessing procedures on the DE2 board which offers great flexibility and portability at the expense of slight penalty in processing time.