Fingerprint Recognition System by Termination Points Using Cascade-Forward Backpropagation Neural Network
Fingerprint authentication belongs to one of the oldest biometric systems. This paper defines a new approach for fingerprint recognition. In this paper only termination points of minutiae are used for authentication. This system matches only the fingerpri
- PDF / 328,490 Bytes
- 9 Pages / 439.37 x 666.142 pts Page_size
- 57 Downloads / 239 Views
Abstract Fingerprint authentication belongs to one of the oldest biometric systems. This paper defines a new approach for fingerprint recognition. In this paper only termination points of minutiae are used for authentication. This system matches only the fingerprint image with database image when there is 100 % match or more than 90 %. Finally, the neural network approach is applied for measurement of neural network performance. The false accept rate and false reject rate are also defined. Keywords Biometric network Matlab
Fingerprint Cascade-forward backpropagation Neural
1 Introduction Biometric recognition is the most oldest and important technology for personal identification [1]. Biometric offers more security than any other traditional method of personal recognition, which is first introduced by Alphonse Bertillon [2]. Biometric authentication is based on physiological and behavioral characteristics. Fingerprint recognition comes in physiological characteristics. The uniqueness of the fingerprint lies on its own characteristic, i.e., minutiae which are unique to everyone [3]. The identical twin does not have the same fingerprint image pattern. The proposed system focuses on termination points that are sufficient for fingerprint authentication. The basic idea of the system is to extract the termination points and Annu Agarwal (&) A.K. Sharma Sarika Khandelwal Computer Science & Engineering, Geetanjali Institute of Technical Studies, Udaipur, India e-mail: [email protected] A.K. Sharma e-mail: [email protected] Sarika Khandelwal e-mail: [email protected] © Springer Science+Business Media Singapore 2016 S.C. Satapathy et al. (eds.), Proceedings of the International Congress on Information and Communication Technology, Advances in Intelligent Systems and Computing 439, DOI 10.1007/978-981-10-0755-2_22
203
204
Annu Agarwal et al.
save it in the database; then the authentication process is applied and the neural network performance is measured. The advantage of using neural network is fast and easy computation. Before the storage of fingerprint image, it is also preprocessed to remove the noise and enhance the image quality. So the database is made with good quality termination points of fingerprint image.
2 Related Work In fingerprint matching, various methods are used for good authentication results. Various researches are carried out which are defined as follows: Mohammed and Nyongesa [4] define fingerprint classification system using fuzzy neural network. This method gave good results. Hsieh and Shing [5] proposed a different method for fingerprint recognition. They used ridge bifurcation for matching process. Their experimental results define that fingerprint minutiae are reliable and robust. Karu and Jain [6] define fingerprint classification system. In this fingerprint is classified into five classes, i.e., right loop, left loop, arch, tented arch, and whorl. In this paper, singular point extraction is defined. The given experiment is invariant to rotation, translation, and small amounts of sca
Data Loading...