An Efficient Singularity Detector Network for Fingerprint Images

Abstract

Singular point related to a fingerprint is a special location that has been used for long to classify the fingerprint in specific pre-defined categories. It is also useful for the alignment of a fingerprint for better matching and comparison. This paper proposes a novel end-to-end deep learning model that incorporate pre-processing, enhancement, segmentation and localization for accurate and efficient retrieval of a singular point. The proposed model has been tested on two databases viz. FVC2002 DB2_A and FPL05K and achieved a Correct Detection Rate of 96.12% and 91.1% respectively which is better than any other state-of-the-art technique.