The characteristics of the friction skin that generates the fingerprint are unique to each person and remain unchanged from the time of formation until decomposition after death. Identification is made on a system based on the reliable extraction of such features from a fingerprint image. Due to the uniqueness and permanence of this features, also called minutiae, the usage of fingerprint for identification purposes has begun in ancient times, bloomed in the late 1800s and early 1900s and is now the biometric most used in everyday life of human kind. However the fingerprint image in hand is not always of good quality, the image may not be clear, distorted, noisy, etc., and this could be due to many factors including acquisition method and environment condition. Since the performance of the system relies on the quality of the features extracted, in order to compensate for these negative factors, the fingerprint image has to go through different stages in order to be properly matched for identification. These stages are image pre-processing, feature extraction, post-processing and finally matching. The aim of this project is to implement a fingerprint identification system tool with Matlab Software in order to be used for teaching and research purposes by evaluating and comparing the performance of two algorithms for each of the main stages of the identification system.
Requirements
Ø Research in the area of Biometrics and Image Processing.
Ø Implement two algorithms for each main stage of the system and evaluation of the performance of each one in comparison with the other.
Ø Implement a GUI to illustrate the performance of the system.
Ø Both the system and the GUI have to be implemented using the Matlab Software.
