An algorithm has been developed to classify various types of fingerprints into 10 classes by using a syntactic method. The image is divided into 16 by 16 segments and directional codes are chosen for each segment. By smoothing directional codes, primi...
An algorithm has been developed to classify various types of fingerprints into 10 classes by using a syntactic method. The image is divided into 16 by 16 segments and directional codes are chosen for each segment. By smoothing directional codes, primitives of a fingerprint are extracted. One dimensional string structures are obtained from their connective relations. As a result of experiments of 50 fingerprints, the classification rates are 94%.