Fingerprint based identification

Acknowledgement

I am greatly indebted to Dr. Yang Xue “B.sc., Msc, PhD, AFSRPSoc MInstP CPhys Csci Senior IEEE ” for assisting me with this project and for valuable input in the organising Project, Constant Encouragement, Consistent Guidance.

I express my sincere thanks to our Module Leader, Prof. Dr. Karim Ouzzane for being a great pillar to support me for all achievements.

I would like to thank Shelly Ahmed ‘CCTM subject Librarian’ for guiding me library resources and subject guide tutorials for this project.

Abstract

     Existing security measures basically depends on the knowledge based approaches like passwords, tokens based approaches used in swipe cards and passports to control physical and virtual spaces. As very common such type of methods are not very secure as cards can be stolen and shared. Moreover, they cannot be differentiating between authorized user and the person having access to the tokens and passwords. There are so many Biometric techniques in this modern world i.e. DNA recognition, face recognition, voice recognition and fingerprint recognition. Fingerprint matching is one of the most famous and unique biometric technique uses to get the main Authentication of users and to store data about people. Fingerprint based identification is the oldest method which is used in so many applications because of every individual is known to be have unique fingerprints [1]. The uniqueness of a fingerprint can be determined by the pattern of the ridges and furrows as well as the minutia points [2]. Fingerprint images are rarely of perfect quality, thus in order to filter images we need to pass image many filters. Reliable extraction of the fingerprint image features and original prints of the furrows and ridges are very challenging problem. There are so many algorithms used for the fingerprint feature extractions and enhancements like co-relation based algorithm, point match based algorithm, phase -correlation algorithm, minutiae matching algorithm and many more. This Research emphasises on diverse emerging fingerprint image comparison techniques. In addition to comparison, this application deduce about the number of angle matched in both print images. A study has been conducted on various soft wares which are associated with the image comparison and fingerprint matching. Neverthless various previously researched software comprises certain limitations in matching fingerprint images and calculating minutiae and angle of orientation. A designed interface is built in order to enable to user to understand and perform matching of the images. This research did not contradict previous researchers; most of my research based on previous researchers theory and previous studies.

Aim:-

     The aim of the proposed research study is to develop fingerprint matching software which would compare two fingerprint images taken by ordinary paper scanner and enable user to know whether the both fingerprint images are matched or non-matched. This software also calculate the angle of orientation and let the user know how many percentage of angle values matched with the other second image and how many minutiae values matched with the second image.

Introduction:

     More than a century has passed, when “Alphonse Bertillon” conceived and then industriously practiced the idea of using body measurements for solving crimes. As his idea was gaining popularity [6] it faded into relative obscurity because of the far most significant and practical discovery of the uniqueness of the human fingerprint [17]. Soon after this discovery many major law departments and researchers embraced the idea of the “booking of the fingerprints” for the criminal records, so that their fingerprints are readily available when they need to verify with the leftover prints on the crime scene, and the identity of the criminal can be determined[5][6]. These agencies sponsored a rigorous study of the fingerprints and developed scientific methods for the visual matching of the fingerprints and strong programming and cultures for the training fingerprint experts, and applied to the art of fingerprint identification for nailing down the perpetrators.

     Despite of the ingenious methods improvised to increase the efficiency of the manual of fingerprints indexing and search, the ever-growing demands on manual fingerprint identification quickly became overwhelming. The manual method of fingerprint indexing resulted in a highly skewed distribution of fingerprints into bins. The most fingerprints fells into a few bins and this resulted in search inefficiencies.

     Most recently, increasing identity frauds has created a growing need for “biometric technology” for positive person identification in a number of non-forensic applications. There are so many questions in “Public and Government sectors” commonly are

1. Is this person authorised to enter this facility.
2. Is this individual entitled to access the privileged information?
3. Is this given service being administered exclusively to the enrolled user?

     To answer these types of questions non-forensic organisations was forced to use fingerprint identification systems. Because biometric systems cannot be easily misplaced, forged or shared. They are more reliable for personal identification then token and password based methods [2].

     The objective of these applications are user convenience (e.g. Money withdrawal without ATM card and PIN), better security (e.g. difficult to forge access), and more efficiency (e.g. lower overhead of computer password access and maintenance).

Fingerprint Based systems Offer several advantages over traditional security measure. These are:

1. Non- Repudiation:-

     With tokens and Password based approaches, the perpetrator can always deny committing the crime that his /her password or Id was stolen and compromised even when confronted with an electronic audit trial. There is no way in which his claim can be verified effectively. This type of problems is known as Problem of deniability or of ‘Repudiation’.

While biometric (Fingerprint) is indefinitely associated with user and hence it cannot be lent or stolen making such repudiation infeasible.

2. Accuracy and security:

     Password based systems are prone to dictionary and brute force attacks. Moreover such systems are as venerable as their weakest password. On the other hand, fingerprint systems require physical presence of the User, therefore cannot be circumvented through a dictionary and brute force style attack. Biometric have also been shown to posses a higher bit strength compared to password based systems [4] and are therefore inherently secure.

3. Screening:

     In screening applications, systems are used to prevent User from assuming multiple identities (e.g. using multiple driving licences and Bank accounts). This requires ensuring that a person has not already enrolled under another assumed identity. Such screening is not possible using traditional authentication mechanisms and fingerprint or biometric systems provide the only available option.

Basics:

     A fingerprint is comprised of ridges and valleys. The fingerprint images are represented by global and local features. The global features include the ridge orientation, ridge spacing and singular points known as core and delta. The singular points are very useful from the classifications perspective [5].The ridges are dark areas of the fingerprints and the valleys are the White area that exists between the ridges and the valleys, as shown in the in the fingerprint image.

The point where a ridge terminate and the point where a single ridge split into two ridges are called “Ridge Ending” and “ Bifurcation” points respectively. A ridge which is shorter than the average ridge length on fingerprint is called “Island (Dot)”. There are so many points and patterns in fingerprint like “core”, “crossover” etc.

     These all points are also known as Minutiae of the fingerprint. The Most commonly used minutiae in fingerprint recognition technologies are “Ridge Endings” and “Bifurcation”, because they can be easily detected by only looking at points that surround them. Minutiae and patterns are very important in the analysis of the fingerprint, as no two different fingerprints shown to be identical. Minutiae based “Fingerprint Matching System ”usually returns the number of matched minutiae on both query and reference fingerprints and uses it to generate similarity scores. Mostly, more matched minutiae generate higher scores. That is when the number of minutiae on both fingerprints are large we can find the impostor fingerprint using the number of minutiae [7].

Objectives of the proposed study:-

     The objective of this study is to study and design and implementation of the ‘Fingerprint Image Matching’. The idea of the project is to study and design of the “USE CASES”, “flow charts”, “sequence diagrams”, “class diagrams”, “charts” and implementation of GUI Model.

Theoretically study of fingerprint matching system’s functions.

• Working of Desktop based fingerprint matching system.
• Working and study of the Image processing Algorithms and Filters.
• To read JPEG scanned images and there related colour values.
• To study about different fingerprint matching techniques and filters to develop code for the project.
• To develop a simple user interface for fingerprints that gives accurate output.
• To Implementation of the real time based tool.

Assumptions and Constraints:-

• The source code is developed in the java environment.
• The main aim of the study is to compare two fingerprint images and calculate angle of orientation values.
• Database is not needed for this project.
• Matching can be done only after uploading both images.
• There is no saving function as saving function is not needed for this tool.

Functional requirements Of the Study:

This project’s primarily specification is to develop software related to the research study on fingerprint matching. Only the software has to be developed for this research. This project is developed by individual with the help and in the guidance of the one appointed supervisor by the University. There is no other involvement in the development of this software. There are various types of resources needs to meet the requirements of the project, which are used for the design and implementation of the project.

1 Hardware Resource:

• Computer System.
• 1 GB hard Disk space.
• 256 MB RAM.
• A scanner to scan images.

The computing lab of London Metropolitan University is well equipped lab with all the facilities required for the environment of study and provides all the Hardware resources and tools which are required for this project. A separate USB pen drive is used to regularly back up the data.

2. Software needs:

• JAVA platform JDK 1.2, which support the robust in build java platform virtual machine.
• J creator LE, software for the environment of the java language platform.
• MS Office 2007, Concept draw office Project management Software, for documentation and Diagram Drawing.

     These facilities are provided by the University’s well equipped computer lab and meet all the need for the project.

3. People :

This people involved in the project for teaching and the guideline is:

• Project Supervisor/Faculty Member : Dr. Yang Xue.

System requirements:-

Hardware Interface:-

No hardware is developed for this system.

Software Interface:-

     The online automatic fingerprint recognition system is very accurate and large system in existing applications. It has an interface the enables the user to verify and identify the related data from the data base with the help of the fingerprint images. This system interface called fingerprint matching system is build a user friendly approach and hence do not find it cumbersome to work on it. ISO quality standards are maintained while developing the software and the interface is associated quite strongly with the quality requirements mentioned in ISO standards. Interface has File, Open and exit Menu option. It also has One text are and one button option to match both fingerprint images and display the result in the text area.

Functional requirements:-

Functional requirements are defined as specific functions, tasks or behaviour the system must support. In term of ISO quality characteristics for evaluation the functional requirements address the quality characteristics of functionality. A functional requirement outlines the needs of the application. [37]

     Functional requirements for the fingerprint matching applications are:-

• The software should be very friendly and should give accurate and quick result.
• User can select fingerprint image from the system. The interface should have a browsing screen so that the user can select file form the various images.
• The selected image should be saved to system memory and keep save the original file till the result is displayed.
• Matching two similar fingerprint images and two different images.
• The user when compare two images, if both are similar it should display a message both images are matched otherwise display both images are not matched.
• The software should have exit button to stop the process.

Non functional requirements:

Non functional requirements are usually some from of constraint or restriction that must be considered when designing the solution. Non functional requirements tend to identify user constraints and system constraints.[37][38].

USABILITY:-

     The image comparison software should be user friendly and easy to understand. Interface should be comfortable so that it will be easy for the user to operate. The tool must be intuitive or easily understood after the tutorials and user guide. Display should be presentable.

CHANGEABILITY:

     The software should be designed in such a way that any future modifications or improvisation can be done with much more functionality.

Efficiency:

     The project must be efficient such that it can select image and display in scaled manner within the GUI. Software should provide reliable result.

Portability:

     The system should be able to run on any platform. This project runs on windows 98/2000/ XP/vista operating system. It should have java virtual machine (JVM) in the operating system and Xinox Software Jcreator LE.

Understandability:

     The fingerprint matching software should enable user to understand whether the software are suitable. The software should be understandable and easy for user to know more about fingerprint matching and make further development in future.[37][38].

Problem Description:

     In current global and competitive environment, information technology plays an important role which results in automation and time saving. There are many types of software which perform fingerprint identification, verification and matching. The main problem in this kind of application is that complexity factor is involved. The interfaces designed in those applications, are very complex and contain highly sophisticated code.

     There is no prior software available that incorporates the feature of matching within the JVM environment, and they keep in standard size in the GUI which looks not impressive. Each time a user wants to match two fingerprint images of his choice, each time the source code would have to be modified. So the Developer should have to care about all these problems and understand the software and user needs.

Patterns of Study:-

This study is organised in the four Chapters to manage each and every part of the process step by step and organise the functions and research in an organised manner. Here shows the chapters and the work done, explained in the chapters.

A. Chapter 1:

Chapter 1, discuss about the basic designing of the project is reviewed and describe about the step by step designing of the project. In this chapter the literature parts shows the study done for the designing and the first step for the understanding and analysis of the project. A study of basic proposed Architecture basic functions of Architectures are discussed for the understanding and working of the project.

B. Chapter 2:

Chapter 2 describe about the various present algorithms, pattern Recognitions and feature extractions of the fingerprint image. These chapters review about filters and the working of the filters.

C. Chapter 3:

Chapter 3 describes about the implementation part of the project and describe about the code used in the implementation. It describes the working pattern of the coding and analysis the basic understanding of the coding.

D. Chapter 4:

Chapter 4, reviews about the testing part of the project and discuss the screen shots of the project taken while testing. It also describes the problems phased during testing and proposed solutions. More it describe about the errors and bugs which will cause the problem while transferring to new proposed system.

1. Project Plan:

     The project has been assigned on the 10 February, 2010. The overall project will be completed till the 4 June 2010. A project plan was drawn on the 2nd week of the Feb. 2010 and worked with the help of Gantt chart to aid smooth project management. The development process of the project is identified as follows:

• Study of existing real world application (like smartcard finger match etc.)For understanding the background and real world working of the application.
• Identify the key area of the concentration and study of the project, moreover needs and environment for the project. A project plan and management chart for better process.
• Study the various algorithms and popularity with in the real world applications and study the main functioning of the algorithms and about the problems faced by them in the past.
• Risk analysis and ethical issues and project proposal report.
• Designing the prototyping methods and Model Building for the look and outer design of the method.
• Implementation of the design and compare with the other application and reviews with the other application.
• Testing the project and making correction according to the needs while facing problems.
• Documentation and Report writing.

The Gantt chart shows about the project plan made for the development and implementation of the project. This specifies resources also who meets the needs of the project.

Schedule summary:

     The schedule of the project is based on this schedule summary of the research study. The schedule of the project is based on the work breakdown structure of the study and pattern of the study. We can measure cost and timing of the project completion. The work breakdown structure is hierarchical chart that represent the tasks, sub tasks and deliverables. This structure is been illustrated in the detail in the Gantt chart. Gantt chart is developed with the help of “Concept Draw office project” (Business and Management Software).

Chapter 1.

Introduction:-

     As I discussed above this chapter discuss the basic designing process of the software and reviews about the step involved in the project management and development. This section encompasses the explanation of the diagrams that exhibits the flow of the data between classes, methods and attributes. Fingerprint matching software is designed to develop user friendly software that provides result very efficiently.

1 Literature Survey:

     Large study but reviews specific literature was done for basic ‘Use case’ designing of the projects which are focused on the project needs.

• Bruegge, Dutoit [2004], explained about the software designing about UML diagrams, about Use cases and structure and step to step configuration of the software. This books review the main concepts and about the structural process of the software and about how to flow from one step to another. [8].
• Springer, Jose [1999], [9] provides details of software management process and methodology of software. A bridge between the real world and the process of project introduced with the help of survey based techniques. The focus is on supporting the development of explicit models and deploying automated support environment. An analysis of core technologies and basic concepts underpinning software process modelling and software process automation, with a special emphasis on the mechanism supported by software evolution are summarised.
• Ray turner [2007], [10], detailed about the methodologies used while in software processing and reviews about the structure charts, functional specification, testing steps.
• Andrew, jeniffer [2005], [11], briefly explains about the life cycles of software project, tools and support. It also discusses the ISO 9000 based eight core principles. Project design maintenance and programming controls are reviewed by this book.
• Deitel, Deitel & sentry [12], provide the full functioning knowledgeable tutorials about programming in java classes. It also discusses the coding platform requirements and about libraries
• Evelyn Stiller and Cathie leblanc, [10][11], details about the development step which are used for software construction under the Conceptualisation view of the system and real world requirements. Moreover functional and non functional requirements can be analysed with the help of this study. It discusses the paradigm and product design key concepts between real world situation with the help of class design and product design process. Evelyn and Cathie, discussed designing and management in parallel views for better results.

System Architecture:

     The system’s general architecture consists of components of fingerprint matching software. This general overview explains the interaction between the User and the Interface. This is representing in the form of simple diagram.

     After analysing the above architecture and working of the system, a working architecture with the method and the classes is drawn, that process the internal working of the system. As per this architecture, initially user loads two images to the image buffer, on which all the operations are performed, such as averaging, filtering and matching.

Methodological Architecture:-

     From this architecture user upload both images to the buffer one by one in the parallel frame and all the operation work in parallel one after second image and provide result.

According to the propose architecture User load both raw images taken by normal scanner and shows both to the main frame in parallel way. Enhance the image with the help of Average method of images to reduce the noise and after that binaries the image in the form of 0 and 1 according to the resolution and intensity of the image. Calculate the angle of orientation through Gabor filter method and extract the features of the both image. With the help of these features, they both features are compared to one by one and calculate the match percentage of the features.

Main Window Design:

The above screen shot is the main window of the fingerprint matching system. This frame appears when we execute the main fingerprint matching software. This window is called the main window known as GUI of the application. This window helps the user to communicate with the application friendly manner and helps for processing all the available options.

     Interface is built in a presentable approach for the user to easily understand the software. It enables the user to perform fingerprint comparison in order to obtain calculated result by the application.

     The fingerprint matching interface of the application consists of File, open and Exit options. This software consists of one button called “MATCH_BUTTON” and one text area which shows all the information about image calculations.

File Menu:-

     There is only one menu option in this software, which further consists of 2 submenu options. File menu consists of two submenus, OPEN and EXIT.

OPEN:

     When user select the OPEN menu a small window appear with the option of File chooser window. File chooser window aids the user to choose the file from desired location. The image should be JPEG image. The purpose of this menu item is to select any image to the interface. The OPEN item works for two times. To select second image to whom User going to Compare user need to press OPEN item again and can load second image without any error. Therefore one item works twice without any obstruct.

EXIT:

     By pressing the EXIT item the application is closed.

TEXT AREA:

     When user load images and perform any action information related to the image is shown in the text area. Like when user load 1st image it shows ‘image 1 is loaded’.

MATCH_BUTTON:

     The function of the button is to do all the calculation after it pressed.

When user clicks the button, the application performs all the calculations and applies filters and calculate angle, compare both image data and produce result in the form of dialog box.

Mind Map Diagram:

     The mind map diagram maps most of related study with the application and the related results. The mind map depicts the related areas and mapping of the mind to concentrate all the related areas of the study. It link all the application related issues and sector of study, problems, selection of the techniques and steps for the future work.

Flow chart Diagram:

     The flow chart diagram depicts the main flow of the application. It shows the background functionality and the processing of the application according to filters and shows the overview of the working.

UML Diagrams:

Overview of UML:

     UML stand for Unified Modeling Language. UML has evolved from the work of ‘GRADY BOOCH’,‘JAMES RUMBAUGH’, IVAR JACOBSON and the relational Software Corporation. This renewed computer scientist fused their respective technologies into a single, standardized model. The Object Management Group (OMG) accepted UML as the standard for Modeling object oriented programs.

     Unified Modeling Language is a standard language that has been designed for the System’s development visualisation, documenting and specifying the requirements of the process. [18][19].

Types of UML Diagrams:

UML basically define nine types of Diagrams:

1. Use case Diagram.
2. Class (Package) Diagram.
3. Sequence Diagram.
4. Collaboration Diagram.
5. State Chart Diagram.
6. Object Diagram.
7. Activity Diagrams.
8. Deployment Diagram.
9. Component Diagram.

Use Case Diagram:

     A use case diagram describes a sequence of actions that provide something of measurable value to an actor and is drawn as a horizontal ellipse. [21]

Class Diagrams:

     Class diagram describe the relationship between classes. It represents the fundamental architecture of the system. This diagram contains the methods and the attributes of the particular class. Attributes are the information stored about an object.

Sequence Diagrams:-

     Sequence diagrams describe interactions among classes in terms of an exchange of messages over time. [19]

Collaboration Diagrams:-

     Collaboration diagrams represent interactions between objects as a series of sequenced messages. Collaboration diagrams describe both the static structures and the dynamic behaviour of a system. [21]

State Chart Diagrams:-

     State chart diagrams describe the dynamic behaviour of an application in response to external stimuli. State chart diagrams are especially useful in modelling reactive objects whose states are triggered by specific events. [20].

Object Diagrams:-

     An object diagram describes the static structure of a system at a particular time. They can be used to test class diagram for accuracy, relationship between the instances and reflect about the multiplicity and roles of the classes. Object diagram uses a subset of the class diagram’s elements. We can draw an object diagram to illustrate a real life example of a class and its attributes relationship. [19]

Activity Diagram:-

     Activity diagram illustrate the dynamic nature of a system by modelling the flow of control from activity to activity and display all the sequences of the activities. Activity diagrams are used to explain all the situations about parallel processing of some activities in the execution time. They represent an operation on some class in the system and about the system’s change in resulting operations. This type of diagrams is used to model business process, internal operations and workflow of the system.

Deployment Diagram:-

     Deployment diagram shows the physical resources in a system, the configuration of the hardware elements (node), software elements, connections and artifacts are mapped onto those nodes. [20].

Component Diagram:-

     Component Diagrams depicts the organisation of physical software components, run-time (binary) code, Source code and executables.

Use Case Diagram for Fingerprint matching system:-

     The use case diagram in this application shows the working pattern of the application. According to this Use case, the user run the application and uploads both raw images to the application. Application stores both images to the JVM memory and process the both images paralally. After running the application shows the main Window to upload raw images taken by Hp normal scanner. Images are taken by normal scanner are considered as normal images, while all the process is done by the application.

Chapter 2

     As according to the pattern of study this section of the research discuss about the fingerprint patterns, feature extraction algorithms, filters and the algorithm and filter used in the application.

Fingerprints and pattern Recognition:-

     Fingerprint is not a separate field. It evolved through interaction and influence of several fields. Fingerprint matching emerged from the recognition of the pattern recognition to forensics. While fingerprints are primarily considered as application of pattern recognition techniques, it has several outstanding differences from conventional classification problems as enumerated below:

1. In conventional pattern classification problems such as ‘Optical character Recognition’ (OCR) recognition, the number of pattern to classify is small compared to the number of pattern with the available fingerprint. Moreover it is very common that only single type of data is generated for each image.
2. The primary task in fingerprint Matching is that of choosing a proper feature recognition representation. Once the feature are carefully chosen, the act of matching the fingerprint are straight forward and commonly employs simple metrics such as Gabor values and theta values are calculated easily. Hence the most challenging aspect of the fingerprint matching image signals processing and image feature extracting.
3. Modalities such as fingerprint, where the image is expressed as minutiae point set do not fall under the category of traditional multi variant and vectorial features commonly used in patter recognition.

Sensors used to get fingerprint images:-

     There are many numbers of sensors and scanners are developed in this market to obtain the fingerprint images.[42] They include:-

Optical sensor:-

• This type of sensor uses a small camera to take image of fingerprint pressed against a clear screen.
• This device capture and sense 3 dimensional surfaces.
• Difficult to fool them with a photograph or other images.

Thermal Sensor:-

• This sensor is temperature based sensor.
• This sensor measure temperature difference between ridges and valleys of fingerprint surface.

Capacitance Sensor:-

• This sensor is electronic capacitance based.
• It measures the pressure difference between ridges and valleys.
• Output is based on electronic charge according to the pressure.

RF Sensors:-

• This sensor is based on radio frequencies and measure frequency related to the fingerprint image.
• It read the frequency of fingerprint from the dermal layer which exists underneath the fingerprint surface.

Ultrasonic Sensors:-

• This sensor can measure fingerprint image beneath the skin.
• Measure frequency, amplitude and charge related to the underneath surface of the skin.

Correlation based Image Defect detection:-

     A brief description regarding the defect detection of the image has been conceptualised by “Toshiyuki” (2006). It was evaluated that the various methodologies which are either model based or non model based using different methods like neural based occlusion detection wavelet transform and ICA and correlation based.

     This paper proposes a new non model based approach to detect image defect, this is based on correlation that does not use learning sample or prior knowledge. In this Approach of defect detection, it is assumed that image description rule is extorted from local region and a defect can be defined as the areas that are not compiled with the rule.

     This description rule of image will be set based on correlation of the image’s local region by the autocorrelation property of image because of image factuality. If this assumption is approved, we can extract the essence of image description for finding the principal components.

Image Processing:-

     Image processing consists of three stages which are Image filtering, Image Enhancement and Image Shaping. In Image Filtering Every Image Consists of pixels in which each pixel consists of three values of colour in different ratios, ranging over 0 to 255 RGB values. These values are referred as three parameters such as red, green and blue as constant figure. For example, return values as RGB of all 0 contains Black and return values of all 255 contains white. In Image Enhancement the black colours mean the ridgelines, they don’t have the same thickness all over the image to overcome this problem of smoothness of the image. From lined image each single black pixel is found out. This draws a filled circle corresponding to the pixel, to smoothen the enhanced image. [44]

     There are many algorithms and filters are proposed by researchers for image processing, basic existing algorithms and filters are studied in this research. Which are as follows:-

Fingerprint matching algorithms and Filters:

     There are many exists algorithms and filters for fingerprint matching systems.

Common Algorithms used in fingerprint process:

1. Minutiae Matching algorithm
2. Correlation matching algorithm.
3. Point Pattern matching algorithm.

1. MINUTIA MATCHING ALGORITHEM:-

     Fingerprint contains a minutiae point, which are made by either ridges ending or ridges bifurcation. In minutiae matching algorithm technique first find the minutiae points and then maps their relative placement on the finger.

     A typical finger has 30 to 60 minutiae points, with the unique location and orientation of there points without any change over a person’s lifetime. The minutiae match the location and orientation of these minutiae points. These are extracted onto a system of co-ordinate during a fingerprint scan and are binary encoded. Temperature difference is measured, primarily not of the fingerprint lines but between the ridges and the valleys of the finger lines. This makes for a robust finger scanning system.

     During the process of the application, System extracts the location and orientation (x, y) of all the minutiae points present in the fingerprint images.

Image Acquisition:-

     The images are obtained in two different ways in existing fingerprint systems:-

• Live scan Print: – A flat bed scanner is user to scan the fingertips and obtain image related to the fingerprint.
• Offline print: – This is the traditional method which is used to obtain the fingerprint by taking the impression on a card /paper in ink and which is later submitted to the system.[1][2][3]

Image Enhancement:-

     This step is a critical step in fingerprint matching systems. In this step minutiae are extracted from the input fingerprint image. This process contains many pre process stages to measure the elements of the fingerprint.

Segmentation and Normalisation:-

     This is the process of extracting the foreground and background of the image and measures them. According to this process the images are divided in small blocks and processed to calculate the variation of the image. As shown below:-

Let’s assume if the size of the block is W x W, then the grey level variation is defined as:-

Vk=1w*wi=0W-1 j=0W-1 2Ii,j-Mk. ……1. [42]

• Where V(K) is the variance for block k,
• I(i,j) s the grey level at pixel (I,j);
• M (k) is the mean grey level value for the block k.

     While in the “Normalisation” process rescales and offset of the image is processed so that the each pixel values should be between 0 and 1(otherwise 0 or 1), because of the normalisation process. An image averaging method is used to normalise the image and process for the accurate results, image averaging is done on 3×3 block of images which is discussed further. [42]

Orientation Estimation:-

     This stage estimates the local orientation of ridges in a fingerprint of a normalized image. It requires three variables which are (a) block sigma which is sigma of the sum of gradients moments, (b) sigma of the derivatives of the image block, (c) smooth orients sigma which is used to smooth the final orientation. In this software Gabor filter based orientation estimation is applied to the fingerprint image and calculated angle of orientation according to the pixel values of the image that is discussed later.

Filtering:

     In the filtering process, filters are used to enhance fingerprint image by using oriented filters, which applied to the normalised, oriented images. In this process images are convolved in the filters and the convolution of a pixel (I, J) in the image requires the corresponding orientation values. A Gabor filter approach is used to filter and calculate orientation of the fingerprint image.[1][2][3][42].

Binarization:-

     The binarization process is the main process to store an image in the form of 0 and 1. This is the most critical task in the fingerprint matching system, because this directly processed on filtered image. If Noise exists in the filtered image and shows false ridge line and valleys in the image that can be counted either 0 or 1, which can change values of FAR and FRR.[2][3]

Feature Extraction:-

     The most important tasks in fingerprint matching are the extraction of the image features. Feature extraction is done by applying a filter of 3×3 or 5×5 matrix on the filtered image; let’s suppose an image is a matrix of 3×3.

• If we study binary image after binarization, and apply 3×3 filter on that, then if the central pixel of the image is with value 1 and include neighbouring pixels in the matrices, and the sum of pixels inside the block is 2, then the central pixel is a termination point.
• If the total sum of the block is 4, after applying above process, then the central pixel is counted as the bifurcation point of the ridges.
• If the central pixel is 1 and the sum of the block is not 2 not even 4 then the block is consider to either continuing ridge and continuing valley of the fingerprint image.[1][2][42][43].

     In this stage we calculate all the termination and bifurcation points of the fingerprint image and store them for matching process. As this stage is depend on binarization process which may be include lots of false minutiae points.

COMPLICATION:

     There are so many complications in these techniques which are:-

1. Each time finger is placed on the scanner, the finger deform differently because of this the location of the minutiae changes and behave differently.
2. Image quality also tends to vary day to day.
3. Sometime ridges may look like as valley, which mistakenly counted as minutiae points and make false data to system.

     Each minutiae may described by a number of attributes include its location in the fingerprint image, orientation, type (ridge ending or bifurcation, a weight based on the quality of the fingerprint image in the neighbourhood of the minutiae and so on. Mostly minutiae base algorithms consider each minutia as a triplet m={x, y, ø}

      Where x and y indicates location of the minutiae in 2 -Dimensional co-ordinates with the angle ø (called minutiae angle).

Suppose T and I represent the Fingerprint images then:

T= {m1,m2,m3,……….,mn}; mi=xi,yi,øi ……….. (1).

Where i= 1, 2, 3… n;

I={m’1, m’2, m’3………………….m’n};

m’i=x’i,y’i,ø’i ……………………………………….. (2).

where i=1,2,3……………………………………..m;

2. Correlation based algorithm:

     Correlation based algorithms enhances the image, repair traces and identifies minutiae points. This algorithm records the portion of the image around the minutiae points. The correlation based analysis of the fingerprints is based on the aligned images where the grey scale intensities are used, the cross correlation operation gives us the similarity percentages of the two images,[14][

     Correlation based algorithm uses ridges ending ridges-bifurcation features

Suggested algorithm requires 5 steps methodology:

1. Segmentation processing.
2. Determination of the ridged orientation
3. RP determination
4. Normalization of the images.
5. Proposed algorithms (correlation based).

1. Segmentation: —

     It is the process of the removing the unnecessary section of the images where the actual fingerprint images is closed. Therefore for the relevant sections of the image, no time is spent for pre-processing steps. IN segmentation process the region outside the fingerprint which doesn’t have any useful information are discarded. To facilitate the extraction of the minutiae, segmentation process is needed.

Variation based segmentation process are calculated by (Jain et al, 1997) shown below. [2][3]

mx,y=1w*wu=x-w/2x+w/2 v=y-w/2v=y+w/2 Iu,v. …….3

sx,y = 1w*w u=x-w/2x+w/2 v=y-w/2y+w/2 2Iu,v-mx,y …

Where w = block size.

mx,y =mean intensity value.

s(x,y) =variation of the block centre at (x, y).

2. Determination of the Ridge Orientation:

     The ridge directions are used for determining the RP. The term orientation image often refers to the determination of local ridge orientation in the FI. The orientation is obtained at each pixel. Following steps are improved in the extraction of the orientation.

1. A block of a given size w*w is centred on a given pixel (I) in the image.

2. The gradient in the x and y direction are computed. [Sober operator can be used to compute the gradient in the x and y direction]

3. The local orientation at each pixel can be computed using the following (Hong et al, 1998) methods:-

vx i,j= u=i-w2i+w2 v=j-w2j+w22dx u,v dy u,v; …………..5.

vy i,j= u=i-w2i+w2 v=j-w2j+w2 dx2u,v dy2u,v ; ……… (6)

Ø i,j=12 tan-1(vy i,j)______________(vx i,j) ………….. (7).

Where Ø i,j is the orientation of each pixel, dx and dy are the gradient in x- directional and y directional respectively.

3. Determining RP. To be able to align most FI a RP is required. The most widely used RP is the core point.

4. Normalization :-

     To reduce the effect of the fingerprint pressure difference is performed. It is a pixel based operation which doesn’t chance the clarity of the ridge and valley track on the fingerprint.

     In normalization process, value of grey level is adjusted, so that it lies within a desired range of values. The desired range of the value is usually a pre specified mean and variance. The fingerprint acquit ion process may result in the fingerprint image exhibiting distortion in grey level values along the ridges and valleys,

     The original images usually have low contrast between the ridges and the valleys. The normalization ensures the contracts as high between the ridges and the valleys, but the ridge structure and the flow pattern are not changed. [14]

3. Point Pattern Matching:-

     The minutiae matching can also be viewed as a point pattern matching. This algorithm matches two point patterns by finding an optimal transformation such that a distance measure for the alignment of the two point pattern under this transformation is minimized, according to the computational complexity, (PPM) can be characterized in 2-Dimensional.[15]

     Point patter matching approaches many studies of known as geometry, Hough transform, relaxation, operation Research solutions, energy minimum etc. [17]. Proposed approaches to PPM matching algorithm as discussed in [15] [17]. There are some steps as follows:

1. Clustering methods.
2. Parameter Decomposition Methods.
3. Relaxation Methods.
4. Bounded Alignment
5. Spectral Graph analysis.
6. Genetic Algorithms.
7. Simulated Annealing.

Proposed Algorithm for the Application:-

     As discussed above there are many algorithms approach to the fingerprint image issues and extracting the specified data related to the image. The most and widely used algorithm is minutae based algorithm. There are many suggestion have been proposed to match the minutiae of the fingerprint image. Most of these algorithms suggestion assume that the images captured by scanner should be of approximately same size. However, fingerprint scanners with a small sensing area are proposed in the market by many organisations. The scanner with the small sensing area 0.5” x 0.7”, which are capture only partial fingerprint image are used to consider the average fingerprint image.

     Matching two fingerprints (in minutiae- based representation) is to find the alignment and correspondences between the minutiae on both print images. For matching regular sized fingerprint images, a brute – force matching solution is not feasible, as number of possible solutions increases exponentially with the number of feature point on the image print [17].

     This application is based on minutiae based algorithm. The guidelines for (visual) matching of fingerprint are quite elaborate. A fingerprint expert often relies on subtle and complex reasoning to argue whether two prints are originated from a single finger or two distinct fingers. For example, an expert can visually localize several rich features of a fingerprint with the remarkable accuracy. These features include the minutiae points of the fingerprint and numerical values of the angle orientation and the number of ridges. An expert can reliably use judgements about scars, complex visual textures and ridge thickness to rule false matches. it is not an exaggerations to state that research in fingerprint matching has been mostly an exercise in imitating the performance of a human fingerprint without access to the many underlying information rich features [2][17].

Noise Reduction:-

     A well know by object oriented and digital signal processor researcher, Mask operation is used on image to remove the noise and smoothen them. A small mask is placed over the pixel of interest and a sum of the products of the mask weights and the corresponding image pixel values is computed. This sum replaces the current values of the pixel and produce new image with less noise.[44]

In this operation we usually takes a block of 3 x 3 and process the block with its neighbour values and calculate the average values of the colour and replace the pixel with the average calculated values. We can represent the product of a mask element and the corresponding image pixel by a single letter as:

Simple neighbourhood matching smoothes the image by getting rid of isolated noise spots. However, it is also have consequences of blurring the image. We see this by considering that overall the operation is just equivalent to taking the image and eight copies displaed in the horizontal, vertical, and diagonal. [3][44].

Sobel Operator:-

     There are many ways to perform image filtering and edge detection of the image. In this study I proposed 2-D spatial gradient Sobel operator to measure the edges of the Image. In this process the Operator calculate the gradient of the image intensity at each point of the image and process on convolving the image with a small integer values and filtered in horizontal and vertical direction of the image.

     Sobel operator uses two 3 x 3 block metrics formulation kernels on the image to calculate approximation of the image, one for the horizontal and one for the vertical changes. If we define sdx and sdy as sobel derivative horizontal wise and sdy as vertical wise, then the computation are as follows:

sdx= -1 -2 -1000 +1+2+1*I and sdy= +10 -1+20-2 +10-1*I

Where I denotes the image pixel data and * denotes the convolution operator. [3][44]

Gabor filter:-

     Jain et Al [2] propose a Gabor filter based approach to filter the fingerprint image called fingerprint codes, for its analogies, which is well known ‘IRISCODE’ introduced by “DAUGMAN”. These codes are used for both further classifying and matching the fingerprints. A centre point is calculated to extract the feature of the print. Simultaneously, a spatial tessellation (into a collection of sectors) of the region of interest I performed, and image is decomposed into a set of components by applying the Gabor filer. Gabor filter with different orientation (0, 45, 90, and 135) is applied to the image and the standard deviation is calculated. [2][3][4].

Performance Evaluation:-

     Unlike password and cryptographic keys, fingerprints have high uncertainty. There is considerable variation between fingerprint images taken at different instances of time, of the same User. Therefore matching of the fingerprint image is always done probabilistic. This is big limitation of the system as compared to exact match required by password or token key based approaches. This lead to two types of errors in fingerprint systems:

1. False Acceptance
2. False Rejection.

1. False Acceptance: A duplicate fingerprint image is sometime matched to original fingerprint image.

2. False Rejection: Sometime a genuine user’s fingerprint image may be rejected to match the fingerprint because of poor quality of fingerprint image, these types of error called ‘false rejection’ error.

METHODOLOGIES:-

Research Methodology:-

This project development is contributed by both qualitative and quantitative approaches for the research methods, which are related and affected with each other.

     Qualitative research main goal is to understand a social or human problem for multiple prospective. Qualitative research is conducted in a natural setting and involves a building a complex and holistic picture of the phenomenon of interest. This research is based on subjective work.

     Quantitative research is based on collection and analysis of data. It establishes a relationship between two or more variables. This technique is widely used in both natural and social sciences. This project has also been developed after carrying out research on the various aspects of the image of the fingerprint and the various techniques being used to compare two images. [3][43].

System Development Engineering Methodology:

     This section discuss about the systems of the fingerprint matching system and the Software development Life cycle of the application. This overview starts with a blank box description of fingerprint matching system and then defines internal component functions. I follow with a reference architecture that shows how internal component functions can be used to support the matching [10][16].

Mostly fingerprints are of two types:

1. One-to-one systems.
2. One-to-many systems.
3. Unique to Unique (Proposed system).

     System which compares a submitted individual file to the existing file called ‘one-to-one’ system. While system which compare large number of file with another large number of files without individuality called ‘one-to-many’ system. But this application is ‘Unique-to-Unique’ system, which compare Unique fingerprint with another Unique fingerprint and decides whether they both matches or not. In practice, some duplication images are considered to be matched, depending upon the size, resolution and false number of minutiae.

     According to the proposed architecture of the system, the components of the system contain the data elements and co-ordinate data which is further used to analysis the performance of the function. According to system’s strategy each file has to be window at each processing unit. The goal is compare the data elements of Image 1 with the Image 2 and display result whether they matched or not. This fingerprint system I decomposed into following sub components:-

• Fingerprint Image Capture: Capture one or more fingerprint images in JPEG format and store them to system.
• Feature Extraction: – Process image 1 and image 2 parallely to extract features and store them to separate attributes for next step process.
• Matcher: – A matcher component is designed to match the both image’s elements.
• Decision Logic: – System combine the result and process the matcher and decide whether the fingerprint matched or not.

Software Development Life Cycle (SDLC) of Application:-

     The SDCL (Software Development life Cycle) is a process of developing a software application within the context of Application Life Cycle Management (ALM). The software development life cycle is basically applied to the problems, which are exists in the application development process. Under SDLC process cycle coding and programming languages are applied to solve the problem issues.[10]

The software Development Life Cycle, term refers to the process of developing methodologies, Models and the Software, which are used to develop any kind of software product, which is directly related to the Information Systems or Computers. SDLC is developed into all kind of software development methodologies, which is used to structure, plan and control the process of developing software or user defined Application.[10][23].

The steps Involved in Software development Life Cycle (SDLC) are:

• Analysis.
• Design.
• Implementation.
• Testing.
• Deployment.
• Maintenance.

Waterfall Model:-

     The waterfall model is also known as classic software life cycle model; and was widely accepted life cycle model which represents the software life cycle using processes and products. The whole process of the software development is divided into different process phases which are requirement specifications, Design of software, Implementation and Testing and Maintenance. Each process transforms a product to produce a new product as output and the new product becomes the input of the next process. [26].

Requirement & Analysis:-

     The software requirements which have to be developed are captured in this model processing phase. Requirements are set of functionalities and Constraints that the end user expects from the system. The requirements are gathered from the end user and are analyzed for their validity and the possibility of incorporating the requirements. The requirements Specification documents are created which serves the purpose of guideline for the next phase of the model.

Software Design Process:-

     In this phase the system design is prepared by studying the functional requirements specifications. The system design specifications serve as input for the next phase of the model. This process phase helps in system requirements, Defining system architecture, and specifying hardware.

Implementation & Unit Testing:-

     In this phase the coding is done depending on the modules and units. Each module is developed according to its functionality. As per, waterfall model the system is developed in small programs called Units. Each Unit is tested according to its functionality. Unit testing mainly verifies the modules and units specifications.

Integration & System Testing:-

     In this phase of waterfall model the software system is divided into units and tested for their functionality which is specified in unit testing phase. Units are divided into integrated into a complete system during integration phase and tested to check all modules and units coordinate between each other. Once the testing process is done it is delivered to the customer.

Maintenance:-

     In this phase the problems related to the system are solved after deployment of the software, that’s why this phase of waterfall Model is never ending Phase. During the post development period of the software, problems arise time to time and all these problems are need to be solved and the software should be implemented to accommodate changes. That’s why in this phase problems related to the system are solved after deployment of the system, and called endless phase. [29][27].

     There are various advantages and disadvantages of this model which are compared below:-

Advantages of waterfall model:-

• This model is very simple and easy to use.
• This model is easy for smaller projects where the requirements are understandable for the user.
• This model is risk free way to develop software systems.
• Each phase is proposed and completed one after the other.

Disadvantages of Waterfall model:-

• The waterfall model does not allow for much iteration.
• Estimating time and cost is difficult for each stage.
• The project is Not Partitioned in phases in flexible way.
• As the requirements are added even after at the end of the phase this will increase the cost of system development and adding more functionality to the application.
• The problem is not solved completely during the phase and in fact many new problems arise. This results on bad structure system as not all problems are solved.

Spiral Model:-

     The spiral model is a System Development and Enhancement Model Used as SDLC model in the information technology, was defined by “Barry Boehm” in his article in 1986. This model of development combines the features of the ‘prototyping model’ and the ‘waterfall model’. The Spiral model is useful mainly for large, expensive and complicated projects. [25].

The working process in the spiral model can be generalized as follows:

1. The idea of the new software system and the requirements of the project are defined in much detail as possible. This usually involves interviewing a number of users representing all the external and internal users and other aspects of the existing system.
2. A first prototyping of the new system is constructed from the preliminary design. This is usually a scaled- down system, and represents and approximation of the characteristics of the final product.
3. A preliminary design is created for the software product.
4. A second prototyping is evolver