A Novel Dynamic Fingerprint Segmentation Method Based on Fuzzy C-Means and Genetic Algorithm

Wu Lei,You Lin

doi:10.1109/access.2020.3011025

Abstract

In automatic fingerprint identification system (AFIS), fingerprint segmentation plays a crucial role in improving detection accuracy and reducing the computation time of feature extraction. With the goal of refining performance of AFIS, in this paper, we investigate a novel dynamic fingerprint segmentation algorithm. The proposed algorithm is based on the existed dynamic image segmentation algorithm using fuzzy c-means (FCM) and genetic algorithm. Specifically, relying on different gray level of histogram and improved post-processing method, we establish a well-performed fingerprint segmentation system. The extensive results from our empirical experiments demonstrate the high performance of our proposed dynamic fingerprint segmentation algorithm, and its better performance than other competing approaches.

Highlights

The development of biometrics recognition technology provides a safer and more convenient way of identity verification for human beings
We investigates a dynamic fingerprint segmentation algorithm based on fuzzy c-means (FCM) and genetic algorithm to improve the accuracy of feature extraction and robustness of automatic fingerprint identification system (AFIS)
3) Numerical experiments showed that our proposed dynamic fingerprint segmentation algorithm can effectively increase the accuracy of fingerprint segmentation in AFIS

Summary

Introduction

The development of biometrics recognition technology provides a safer and more convenient way of identity verification for human beings. The pattern of fingerprint does not change during the life of an individual except the accidents such as cuts and bruises on the fingertips. Owning to this property, fingerprint becomes an attractive biometric identifier and plays a critical role in user authorization. There are two distinct parts in a fingerprint image, i.e., background and foreground. The background area is usually overlapped by complex non-fingerprint noise, which may increase the probability of classification mismatch. The foreground area, which contains effective ridge information, is originated from the contact of a fingertip with the sensor. How to extract the foreground area (i.e., fingerprint segmentation) from a fingerprint image becomes a decisive factor to enhance fingerprint identification accuracy

Objectives

Methods

Findings

Conclusion