FEATURE EXTRACTION TECHNIQUES FOR FINGERPRINT RECOGNITION

Abstract

The feature extraction involves the process of converting original data into a set of numerical features that can be 
processed by machine learning algorithms while conserving the significant information from the original data. It's a 
key phase in data preprocessing, specially for intricate datasets, as it can make simpler the data, decrease 
dimensionality, and improve model performance. In fingerprint recognition systems, the feature extraction remains a 
very important stage because it reduces the complexity of fingerprint data, enables effective and efficient matching, 
and enhances the overall performance and accuracy of the system. This process transforms raw fingerprint images 
into a set of distinctive and stable features, such as minutiae points (ridge endings and bifurcations) and global 
patterns, which are then used for identification and verification. The study conducts a literature on feature extraction 
techniques for fingerprint recognition system. Different algorithms were mentioned with associated challenges and 
further future areas of improvement were suggested.