An Efficient Automatic Overlapped Fingerprint Identification and Recognition Using ANFIS Classifier

Abstract

The Automatic Fingerprint Recognition System plays an important role in forensics and law enforcement applications. The objective of the proposed system in the current study is to identify and separate overlapped fingerprint images automatically using an Adaptive Neuro Fuzzy Inference System (ANFIS) Classifier. There are various issues that have been identified, which need to be addressed to develop the scope of light-out fingerprint recognition system. The latent fingerprint images can be overlapped in crime scenes. During investigations, there are several possibilities for acquiring damaged or overlapped fingerprint images. The proposed system analyzes and identifies the overlapped images using an ANFIS Classifier. This paper also proposes a novel algorithm for the separation of overlapped images. The proposed work is designed to retrieve fast and accurate data using fingerprint identification for the overlapped images. Extensive experiments are performed on the FVC 2006 DB1-A, DB2-A, NIST SD27 and SLF databases. The experimental results are highly promising and outperform the previous systems in identifying the overlapped images. Our proposed system separates those overlapped fingerprints more accurately and robustly. The achieved results confirmed that the proposed automatic fingerprint recognition system has higher possibility of overlapped fingerprint detection.

Performances Enhancement of Fingerprint Recognition System Using Classifiers

“…A novel algorithm was proposed for the separation of overlapped images using the ANFIS model, the overlapped images are identified automatically and classified. The experimental results show that the proposed method achieves a separation rate of 92.4% for the recognition of overlapped images on all overlapped fingerprint images in the SLF, NIST, and FVC datasets, however giving the classification rate of 88.6%, with the accuracy of 92% which is comparably low [4].…”

Deep learning‐based approach to latent overlapped fingerprints mask segmentation

“…The most significant latent overlapped fingerprint separation approaches proposed during the past few years [13][14][15][16][17][18][19] include manual segmentation of component fingerprints' ROIs, while the approach proposed in [12] requires manual segmentation of three different regions -background, single, and overlapped regions.…”

“…Very few researchers have worked on single latent fingerprints segmentation [1][2][3][4][5][6][7][8][9][10], and there are, to the best of our knowledge, no published approaches for overlapped fingerprints ROIs segmentation. Manual segmentation is still standard in most significant overlapped fingerprints' separation approaches [12][13][14][15][16][17][18][19].…”

“…A number of methods for separation of overlapped latent fingerprints have been proposed in the literature (see [11] for a recently published comparative analysis of latent overlapped fingerprint separation approaches). In all of the overlapped latent fingerprints separation approaches [12][13][14][15][16][17][18][19] reviewed in [11], the (mask) segmentation step is performed manually.…”

“…The most important impact of the proposed approach is that it offers an opportunity for the development of a fully automated fingerprint separation pipeline (i.e. without requiring manual mask segmentation -which is a requirement of all published approaches [12][13][14][15][16][17][18][19]).…”