Graph based K-nearest neighbor minutiae clustering for fingerprint recognition

The graph is an efficient data structure to represent multi-dimensional data and their complex relations. Pattern matching and data mining are the two important fields of computer science. Pattern matching finds a particular pattern in the given input where as data mining deals with selecting specific data from the huge databases. This work contributes towards the combination of graph theory, pattern recognition and graph based databases. A variety of graph based techniques have been proposed as a powerful tool for pattern representation and classification in the past years. For a longer time graphs remained computationally expensive tool. But recently the graph based structural pattern recognition and image processing is becoming popular. The computational complexity of the graph based methods is becoming feasible due to high end new generations of the computers and the research advancements. In this work we have implemented graph based fingerprint recognition algorithm. The fingerprints are represented as attributed relational graphs. In the pattern recognition phase graph matching is applied. This study focuses on the clustering of graph databases prior to graph matching. When the structural feature set size of the data grows longer, graph matching becomes expensive. The clustering of graph databases drastically reduce the graph matching candidates.