Convolutional Neural Network Approach for Precise Fingerprint Recognition

Fingerprint recognition is a technique which is used for identifying people automatically. This technique is one of the most effective and popular biometric techniques. The effectiveness of this technique is depends on the quality of the fingerprint which was given as input. If the quality of fingerprint is noise free then it is possible to detect a reliable set of features otherwise it is not possible. The name of the dataset used in this research is Sokoto Coventry Fingerprint Dataset. Original data and altered data from the dataset have been used. There have several algorithms for fingerprint recognition which is offered by the recent advances pattern recognition and machine learning. Convolutional Neural Network (CNN) is one of the algorithms for fingerprint recognition. Image processing speed is very high in this algorithm. In this paper, the focus is on recognizing fingerprints with the use of CNN. It have a very high performance and for that reason this algorithm is used for classification and recognition of image. It follows a hierarchical model for build a network just like a funnel and gives a layer that are fully connected. In this network neurons are connected together where the output is processed. This method can achieve about 99% accuracy of recognition if the input data set is noise free and train the data set in a right way.