Multiple Criteria Decision Analysis Based Overlapped Latent Fingerprint Recognition System Using fuzzy Sets

Latent fingerprints have attracted considerable attention from researchers in the fields of forensics and law enforcement applications. Public demand for these applications may be the driving force behind further progress in biometrics research. Although great effort has been taken to devise algorithms for overlapped latent fingerprint classification system, there are still many challenging problems involved in fingerprint classification systems. Most of the fingerprint-based applications will prolong with fingerprint recognition because of its proven performance, the existence of large databases and the availability of the fingerprint devices with minimum cost. There are various issues that need to be addressed to develop fingerprint classification system. In this connection, there are some designing challenges such as nonlinear distortion, low-quality image, segmentation, sensor noise, skin conditions, overlapping, inter-class similarity, intra-class variations and template ageing. In crime scenes, the latent images can be overlapped with some background images or more number of fingerprint images from same person or different person. An overlapped fingerprint image should be processed for fingerprint classification. This proposed recognition system suggests multiple criteria decision analysis technique to assess the overlapped latent fingerprints. The proposed multiple criteria such as first-order, second-order and third-order features are used to classify the overlapped latent fingerprints. The proposed system designs the novel classification system for overlapped latent fingerprint images using ANFIS classifier. Extensive experiments are performed on the simultaneous latent fingerprint databases, and National Institute of Standards and Technology-Special Database 27, Fingerprint Verification Competition 2006 Database1-A and Database2-A databases. The planned work enables accurate and fast data retrieval by using one-to-N fingerprint classification for overlapped latent images. The experimental results are highly promising, and they outperform the existing systems in classifying overlapped images. The performance of adaptive neuro fuzzy inference system classifier is evaluated by applying the k-fold cross-validation technique. The outcome of the work shows that the overlapped fingerprint is classified in a successful manner, and the results are compared with Bayes, SVM and MLP classifiers. The obtained results show that the proposed system achieves a better classification rate of 90.66% with 5 s, 86.66% with 12 s compared to the existing system.