Вы находитесь на странице: 1из 5


VOLUME 3 ISSUE 1 JANUARY 2015 - ISSN: 2349 - 9303

A Survey Based on Fingerprint Matching System

K. Lavanya 1

1 Department of Computer Science & Engineering, Bannari Amman Institute of Technology, Anna University,

M.Krishnamoorthi 2

2 Department of Computer Science & Engineering, Bannari Amman Institute of Technology, Anna University, krishnamoorthim@bitsathy.ac.in

Abstract Fingerprint is one of the biometric features mostly used for identification and verification. Latent fingerprints are conventionally recovered coming in to existence of crime scenes and are analyzed with active databases of well-known fingerprints for finding criminals. A bulk of matching algorithms with distant uniqueness has been developed in modern years and the algorithms are depending up on minutiae features. The detection of accepted systems tries to find which fingerprint in a database matches the fingerprint needs the matching of its minutiae against the input fingerprint. Since the detection complexity are more minutiae of other fingerprints. Therefore, fingerprint matching system is a higher than verification and detection systems. This paper discussed about the various novel techniques like Minutia Cylinder Code (MCC) algorithm, Minutia score matching and Graphic Processing Unit (GPU). The feature extraction anywhere in the extracted features is sovereign of shift and rotation of the fingerprint. Meanwhile, the matching operation is performed much more easily and higher accuracy.

Index TermsGraphic processing unit, latent fingerprints, minutiae, minutia cylinder code and minutia score matching.


A fingerprint is the replication of a fingertip epidermis, produced when a finger is pressured

against a plain surface. The most apparent structural characteristic of a fingerprint is a pattern of ridges and valleys; in a fingerprint pattern, ridges are dark, though valleys are bright. Ridges and valleys are frequently run in parallel and they will bifurcate as well as terminate.

in parallel and they will bifurcate as well as terminate. Fig 1: Ridges & Valley in

Fig 1: Ridges & Valley in Fingerprint


When analyzed at the global level, the fingerprint pattern explains about one or more regions where the ridge lines which are characterized by high curvature, frequent and termination. These regions are called singularities or singular regions which shall be divided into three types: loop, delta, and whorl.

regions which shall be divided into three types: loop, delta, and whorl. Fig 2: Singular regions

Fig 2: Singular regions in Fingerprint



VOLUME 3 ISSUE 1 JANUARY 2015 - ISSN: 2349 - 9303

At the local level, other important features, called minutiae which can be identified in the fingerprint patterns. Minutia refers to dissimilar ways that the ridges can be disconnected. For example, a ridge can unexpectedly come in to an end called termination, or can divide into two ridges. It is a specific point in the fingerprint image. When the fingerprint is obtained from the person, the numbers of minutiae are recorded. Graphical Processing Unit (GPU) is “a single chip processor and used to manage the performance of video and 3D functions”. Some of the GPU typical features are 2-D or 3-D graphics, rendering polygons and texture mapping. The very first GPU was developed by NVidia in 1999 and it is called as GeForce 256. The GeForce 256 model can process up to 10 million polygons per second and have more than 22 million transistors. This model was a single-chip processor with integrated transform and BitBLT support. Most GPUs use their transistors for 3-D computer graphics. Anyhow some have increased memory for checking vertices, such as geographic information system (GIS) applications. Some of the other modern technology supports programmable shaders executes textures, mathematical vertices and explicit color formats. Some applications like computer-aided design (CAD) can process over 200 billion operations per second and transfer up to 17 million polygons per second. More scientists and engineers use GPUs for more in-depth calculated studies utilizing vector and matrix features.


Nearly 70% of consumers’ worldwide support using biometrics systems, such as fingerprints or voice recognition, administered by a trusted organization (bank, healthcare provider or government organization) as a way to verify an individual's identity. Biometrics consists of automated methods of recognizing a person based on unique physical characteristic. Several types of biometric system, but in some of different application, hold at least one similarity: the biometric must be suited upon a distinguishable human attribute such as an individual’s fingerprint pattern.

Nowadays fingerprint devices are by far the most popular form of biometric security used, with a wide variety of systems on the market intended for general and bulk market usage. Long gone are the massive fingerprint scanners; now a fingerprint scanning device can be small.


A fingerprint is made up of a pattern of ridges and

furrows as well as characteristics that occur at Minutiae points Fingerprint scanning essentially provides an identification of a person based on the acquisition and recognition of those unique patterns and ridges in a fingerprint. The actual fingerprint identification process will change slightly between products and systems. The basis of identification, nevertheless, is nearly the same. Some of the standard systems are comprised of a sensor for scanning a fingerprint and a processor which stores the fingerprint database and software which compares and matches the fingerprint to the predefined database. Inside the database, a fingerprint is usually matched to a reference number, or ID number, and then it is matched

to a person's name or account. In case of security the

fingerprint match is generally used to allow or disallow access, but nowadays this can also be used for something as simple as a time clock or payroll access. In large government organizations and corporations, biometrics plays a important role in employee identification and security. Additionally some data centers have moved on the bandwagon and have implemented biometric scanners to enhances remote access and management by adding another layer of network security for system administrators. Unfortunately the cost of invoking fingerprint and other biometric security scanning in data centers is more expensive, and many centers still entrust on ID badges while waiting until biometric technology becomes a little more pocket-book friendly. Today companies have realized that fingerprint scanning is an efficient means of security. While the cost of executing biometric scanners in larger organizations and data centers is still costly and finding several fingerprint scanning devices which would fit

into the budget of many small offices and home users.

1.2 Fingerprint Segmentation

Before extracting the feature of a fingerprint, it is important to separate the fingerprint regions (presence of ridges) from the background[1]. This limits the region to be processed and therefore reduces the processing time and false feature extraction. A correct segmentation may be, in some cases, very difficult, especially in poor quality fingerprint image or noisy images, such as presence of latents.



VOLUME 3 ISSUE 1 JANUARY 2015 - ISSN: 2349 - 9303

The same information used for quality extraction, such as contrast, ridge orientation and ridge frequency can be used for the segmentation or inclusive the quantified region quality may be used directly by considering as background the regions with quality below some threshold. Normally, the segmentation are also computed by block in the same way as the quality extraction.

1.3 Fingerprint Enhancement

The quality of the fingerprint image is determined by many factors which sometime may be difficult to control; therefore a fingerprint system must be able to handle also the image of medium and low quality (recoverable). In some cases it is possible to improve significantly the image quality by applying some image enhancement technique. The main purpose of such procedure is to enhance the image by improving the clarity of ridge structure or increasing the consistence of the ridge orientation. In noisy regions, it is difficult to define a common orientation of the ridges. The process of enhancing the image before the feature extraction is also called pre-processing. Here we show two technique of pre-processing fingerprint images.

The first consist in a simple normalization and the second uses Fourier transformation (FFT). There a lot of other more complex techniques and there still place for future researches on this topic. The enhancement may be useful for the following cases like connect broken ridges, eliminate noises between the ridges and improving the ridge contrast.

1.4 Minutia Extraction

There has been proposed many methods for the minutia extraction, the traditional method consist of the following steps. a) Binarization: This process consist in converting the gray scale image in binary image, i.e, the intensity of the image has only two value: black, representing the ridges, and white, representing the valleys and the background. A simple method to binarize is to use a global threshold value, however, it is not well suited for noisy images, a more robust method consist of using some rectangular mask, rotate according the orientation of the ridges.


b) Thinning: The objective of thinning is to find the ridges of one pixel width. The process consists in performing successive erosions until a set of connected lines of unit-width is reached. These lines are also called skeletons. An important property of thinning is the preservation of the connectivity and topology which however can lead to generation of small bifurcation artifacts and consequently to detection of false minutiae. Therefore some procedure aiming the elimination of these artifacts must be performed after the thinning. c) Minutiae Detection: From the binary thinned image, the minutia are detected by using 3x3 pattern masks. Samples of masks used for identifying the ridge ending and bifurcations point are shown in the figure below. Although the process seems to be simple, it is necessary to consider the elimination of false detected minutiae[2].

After a successful extraction of minutiae, they are stored in a template, which may contain the minutia position (x,y), minutia direction (angle), minutia type (bifurcation or termination), and in some case the minutia quality may be considered. During the enrollment the extracted template are stored in the database and will be used in the matching process as reference template or database template. During the verification or identification, the extracted minutia are also stored in a template and are used as query template during the matching.







The fundamental fingerprint biometric processes are:

a) Enrollment scans: The user’s fingerprints are scanned and associated with that user’s identity in the system. It is supervised process to allow for preventing false identity creation and propagation. In an enterprise scenario, enrollment would be done at the time a person becomes employed and only needs to be done once.

b) Template creation and storage: A biometric template is created from biometric features derived from the scanned fingerprint. The enrollment template becomes the fingerprint biometric record for the user. In some solutions, the fingerprint scan itself may also be stored. See below figure



VOLUME 3 ISSUE 1 JANUARY 2015 - ISSN: 2349 - 9303

VOLUME 3 ISSUE 1 – JANUARY 2015 - ISSN: 2349 - 9303 Fig 3: Biometric Template

Fig 3: Biometric Template Creation

c) Live scan: Each time a person requests access to the system, a live scan of the fingerprint is made and a live template is derived from that scan. The scanner may also perform a liveness measurement, which can distinguish between an artificial copy of the fingerprint and a “live” finger and thus validate the authenticity of the fingerprint. d) Automated matching: The live template is compared to a specific enrollment template and a matching score is generated.


Minutia Cylinder-Code (MCC) algorithm accomplished with local structure for every minutia. This structure encodes spatial and directional connections between the minutia and its neighborhood and can be represented without difficult as a cylinder whose base and height are compared to the spatial and fractional facts. The weaknesses were forwarded by recommending local minutiae matching method. Local minutiae are distinguished by attributes which are invariant with respect to the global transformations like translation and rotation. Matching fingerprints placed on local minutiae arrangements relaxes global spatial, which are highly different; and accordingly reduces the amount of information available for fussy fingerprints. Local minutiae system can be classified into nearest neighbor-based and fixed radius-based. The neighbors of the central minutiae, K which is spatially closest minutiae. These points to fixed-length descriptors which will be matched very conveniently. The descriptor length variables are depends on the local minutiae density and leads to a more complex local matching. It is more sophisticated against missing and false minutiae. Two drawbacks are the absolute encoding of radial angles, and the missing directional difference between the central minutia and the neighboring ones.


The technique proposed by Feng [6] does not endure from the listed disadvantages and can be considered as a new fixed-radius local matching algorithm. The MCC [2] introduces a novel minutiae- only local representation aimed at combining the both advantages of neighbor-based which matches minutiae from two fingerprints with similar neighbor and fixed- radius structures uses the information of all the minutiae at once.

3.2 Minutia Score Matching

RAVI. J and K. B. RAJA [5] presented a paper using Fingerprint Recognition using Minutia Score Matching method (FRMSM). The pre-processing the original fingerprint involves image binarization, ridge ending, and noise removal. The Minutia Score Matching method is used for fingerprint identification to correspond the minutia points. The proposed method FRMSM gives better False Matching Ratio (FMR) values. The need is to identify a person for security. The block diagram of FRMSM determines the match of tested fingerprint with the template database using Minutia Matching Score algorithm.

of tested fingerprint with the template database using Minutia Matching Score algorithm. Fig 4: Block diagram

Fig 4: Block diagram of FRMSM



VOLUME 3 ISSUE 1 JANUARY 2015 - ISSN: 2349 - 9303


Fingerprint Image: The input fingerprint image is the gray scale image of the person’s fingerprint; where the minutiae points are the locations and the ridge becomes discontinuous.


Binarization: The use of Binarization in this process is to convert a gray scale image into a binary image i.e. 0 or 1 by fixing the threshold value.


Block Filter: The binarized image is thinned using Block Filter to minimize the thickness of all ridge lines. It will not change the location and orientation of minutiae points.


Minutiae Extraction: The minutiae location and the angles are derived after minutiae extraction.


Minutiae Matching: To compare with the input fingerprint data with the expected data, so then minutiae matching technique is used.


Matching Score: It achieves the better False Matching Ratio


It is a processor optimized for 2D/ 3D graphics and video. It provides real- time visual interaction with the computed objects. In 2007, Compute Unified Device Architecture (CUDA) was invented. It is a parallel computing platform and programming model created by NVIDIA and implemented by the graphics processing units (GPUs)


In this paper, we have discussed about the various approaches and techniques involved in the Fingerprint matching system. This paper presents the thorough information on the subject of fingerprint biometrics and expressly alert on techniques that overcome the disadvantages of fingerprint biometric system. Fingerprint biometric system is used most of the fields except chemical industries since the finger print of individual’s people working in Chemical industries are habitually affected. It wrap up that Finger print biometrics is solitary of the efficient, protected, cost efficient, ease to utilize technologies for consumer authentication and according to our investigation almost all drawbacks are prevail over in fingerprint biometric system.



[1] M. Jawalkar, V. Sarode and M. Ghonge, “A Survey based Accomplishment Techniques for Biometric Finger Print Matching System,” J. computer applications, 2014. [2] Pablo David Gutiérrez, Miguel Lastra, Francisco Herrera, and José Manuel Benítez “A High Performance Fingerprint Matching System for Large Databases Based on GPU,” IEEE Trans. Information Forensics And Security, Vol. 9, No. 1, January 2014. [3] M. Schatz, C. Trapnell, A. Delcher, and A. Varshney, “High-throughput sequence alignment using graphics processing units,” BMC Bioinformatics, vol.8, December 2007. [4] R. Cappelli, M. Ferrara, and D. Maltoni, “Minutia Cylinder-Code: A New Representation and Matching Technique for Fingerprint Recognition,IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 32, No. 12, December 2010. [5] X. Jiang and W.Y. Yau, “Fingerprint Minutiae Matching Based on the Local and Global Structures,” Proc. Int’l Conf. Pattern Recognition, vol. 2, pp. 6038-6041, 2000. [6] J. Feng, “Combining Minutiae Descriptors for Fingerprint Matching,” Pattern Recognition, vol. 41, pp. 342-352, January 2008. [7] J. Ravi, K. B. Raja and K.R Venugopal “Fingerprint Recognition Using Minutia Score Matching,” J. Engineering & Technology, vol.1, No.2, 2009. [8] G.Sambasiva Rao, C. NagaRaju, L. S. S. Reddy and E. V. Prasad, “A Novel Fingerprints Identification System Based on the Edge Detection,” J. Computer Science and Network Security, vol. 2, No 8, December 2008. [9] D. Maltoni, D. Maio, A. Jain, and S. Prabhakar, “Handbook of Fingerprint Recognition,” NY, USA:

Springer-Verlag, 2009, (second edition).