Enhancing Fingerprint Image Recognition Algorithm Using Fractional Derivative Filters

Abstract

One of the most important steps in recognizing fingerprint is accurate feature extraction of the input image. To enhance the accuracy of fingerprint recognition, an algorithm using fractional derivatives is proposed in this paper. The proposed algorithm uses the definitions of fractional derivatives Riemann-Liouville (R-L) and Grunwald-Letnikov (G-L) in two sections of direction estimation and image enhancement for the first time. Based on it, new mask of fractional derivative Gabor filter is calculated. The proposed fractional derivative-based method enhances the image quality. This method enhances the structure of ridges and grooves of fingerprint, using fractional derivatives. The efficiency of the proposed method is studied in images of FVC2004 (DB1, DB2, DB3 and DB4) database and the results are evaluated using the criteria including entropy, average gradient, and edge intensity. Also, performance of the proposed method is compared with other technical methods such as Gabor filter. Based on the obtained results from the tests, the method is able to enhance the quality of fingerprint images significantly.

Fractional Derivatives Application to Image Fusion Problems

“…The application of this calculus in image processing is also an issue that has been intensively researched in recent years. The use of some models has shown that satisfactory results have been obtained, for example, in medical image processing [ 21 ] or photo-based fingerprint recognition [ 20 ]. This calculus is used in problems where there is a memory effect since having memory is a feature of fractional order operators [ 21 , 25 ].…”

“…Recently, among image fusion methods, the emergence of algorithms based on fractional differential calculus can be observed [ 15 , 16 , 17 , 18 , 19 ]. They are applied in such problems as fingerprint detection [ 20 ], exposing of essential elements in medical images [ 21 ], brain photo analysis [ 22 ], elimination of noise in images (improved image quality) [ 23 ], or contrast enhancement [ 24 ]. As the fractional derivative started to be used relatively recently, new applications are still being found.…”

A Review of Recent Advances in Fractional-Order Sensing and Filtering Techniques

“…Experimental data based on well-known biomedical databases prove the efficacy of the proposed algorithm as approximately 95%, showing a significant improvement than other methods. Another interesting application uses Grünwald–Letnikov to develop a fractional-order image filter for digital fingerprint identification [ 153 ].…”

An Effective Denoising and Enhancement Strategy for Medical Image Using Rl-Gl-Caputo Method

“…In this work, to improve the detection of arteries on coronary angiograms a variation in an algorithm. [3] proposed algorithm utilizes the fractional derivatives (R-L) and (G-L) definitions in the estimation of two direction areas and enhancement of image in the course of first run. As per this, new Gabor filter mask fractional derivative is intended.…”

“…The effectual means of making a convolution mask which is fractional-dependent will be based on image denoising and image enhancement mechanism which is capable of recognizing edges quite significantly in detail [1]. The fractional derivatives benefits are obvious in engineering correspondence that covers automatic manipulate, biomedical programs, finite impulse reaction filter designs, and in lots of other fields [2,3]. Noise is signified as any unwanted signal which in turn contaminates an image.…”