Design And Implementation Of An ATM Fingerprint Authentication System

ABSTRACT

Bio metrics technology is rapidly progressing and offers attractive opportunities. In recent years, bio metric authentication has grown in popularity as a means of personal identification in ATM authentication systems. The prominent bio metric methods that may be used for authentication include fingerprint, palm print, hand print, face recognition, speech recognition, dental and eye bio metrics. In this paper, a microcontroller based prototype of ATM cash box access system using fingerprint sensor module is implemented. An 8-bit PIC16F877A microcontroller developed by Microchip Technology is used in the system. The necessary software is written in Embedded ‘C’ and the system is tested.

CHAPTER ONE

• INTRODUCTION

Bio metrics  are  automated  methods  of  recognizing  a  person  based  on  a  physiological  or  behavioral characteristic. Bio-metric-based solutions are able to provide for confidential financial transactions and personal
data privacy. The various features used are face, fingerprints, hand geometry, handwriting, iris, retina, vein and
voice [1].  Fingerprinting  or  finger-scanning  technologies  are  the  oldest  of  the  bio metric  sciences  and  utilize
distinctive features of the fingerprint to identify or verify the identity of individuals. Finger-scan technology is
the most commonly deployed bio metric technology, used in a broad range of physical access and logical access applications. All fingerprints have unique characteristics and patterns. A normal fingerprint pattern is made up
of lines and spaces. These lines are called ridges while the spaces between the ridges are called valleys. It is
through  the  pattern  of  these  ridges  and  valleys  that  a  unique  fingerprint  is  matched  for  verification  and authorization. These unique fingerprint traits are termed “minutiae” and comparisons are made based on these
traits [2]. On average, a typical live scan produces 40 “minutiae”. The Federal Bureau of Investigation (FBI) has reported that no more than 8 common minutiae can be shared by two individuals.

There  are  five  stages  involved  in  finger-scan  verification  and  identification.  Fingerprint  (FP)  image acquisition,  image  processing,  and  location  of distinctive  characteristics,  template  creation  and  template matching  [3].  A  scanner  takes  a  mathematical  snapshot  of  a  user’s  unique  biological  traits.  This  snapshot  is

saved  in  a  fingerprint  database  as  a  minutiae  file.  The  first  challenge  facing  a  finger-scanning  system  is  to acquire high-quality image of a fingerprint. The standard for forensic-quality  finger printing is  images of 500 dots per inch (DPI). Image acquisition can be a major challenge for finger-scan developers, since the quality of print differs from person to person and from finger to finger. Some populations are more likely than others to have  faint  or  difficult-to-acquire  fingerprints,  whether  due  to  wear  or  tear  or  physiological  traits.  Taking  an image in the cold weather also can have an affect. Oils in the finger help produce a better print. In cold weather, these oils naturally dry up. Pressing harder on the platen (the surface on which the finger is placed, also known
as a scanner) can help in this case. Image processing is the process of converting the finger image into a usable format. This results in a series of thick black ridges (the raised part of the fingerprint) contrasted to white valleys. At this stage, image features are detected and enhanced for verification against the stored minutia file. Image enhancement is used to reduce any distortion of the fingerprint caused by dirt, cuts, scars, sweat and dry skin [3]. The next stage in the fingerprint process is to locate distinctive characteristics.  There is a good deal of information on the average fingerprint and this information tends to remain stable throughout one‟s life. Fingerprint ridges and valleys form distinctive patterns, such  as  swirls,  loops,  and  arches.  Most fingerprints have  a  core,  a  central  point  around which swirls, loops, or arches are curved. These ridges and valleys are characterized by irregularities known as minutiae,  the  distinctive  feature  upon  which  finger-scanning  technologies  are  based.  Many  types  of  minutiae
exits, a common one being ridge endings and bifurcation, which is the point at which one ridge divides into two.

A typical  finger-scan  may  produce  between  15  and  20  minutiae.  A  template  is  then  created.  This  is accomplished  by  mapping  minutiae  and  filtering  out  distortions  and  false  minutiae.  For  example,  anomalies
caused by scars, sweat, or dirt can appear as minutiae. False minutiae must be filtered out before a template is created and is supported differently with vendor specific proprietary algorithms. The tricky part is comparing an
enrollment template to a verification template. Positions of a minutia point may change by a  few pixels, some minutiae  will differ  from the enrollment  template, and  false  minutiae  may be seen as real. Many finger-scan
systems  use  a  smaller  portion  of  the  scanned  image  for  matching  purposes. One  benefit  of  reducing  the comparison area is that there is less chance of false minutiae information, which would confuse the matching
process and create errors.

1.1                                           BACKGROUND OF THE STUDY

A bio metric system is essentially a pattern recognition system that operates by acquiring bio metric data from an individual, extracting a feature vector from the acquired data, comparing this feature vector from the database feature vector. Person authentication has always been an attractive goal in computer vision. Authentication systems based on human characteristics such as face, finger, iris and voice are known Bio metrics systems. The basis of every bio metric system is to get the input image and generate prominent feature vectors like color, texture, etc.

Today, bio metric recognition is a common and reliable way to authenticate the identity of a living person based on physiological or behavioral characteristics. A physiological characteristic is relatively stable physical characteristics, such as fingerprint, iris pattern, facial feature, hand silhouette, etc. This kind of measurement is basically unchanging and unalterable without significant duress. A behavioral characteristic is more a reflection of an individual’s psychological makeup as signature, speech pattern, or how one types at a keyboard.

The degree of intro-personal variation in a physical characteristic is smaller than a behavioral characteristic. For examples, a signature is influenced by both controllable actions and less psychological factors, and speech pattern is influenced by current emotional state, whereas fingerprint template is independent. Nevertheless all physiology-based bio metrics don’t offer satisfactory recognition rates (false acceptance and/or false reject rates, respectively referenced as FAR and FRR). The automated personal identity authentication systems based on iris recognition are reputed to be the most reliable we consider that the probability of finding two people with identical iris pattern is almost zero. That’s why iris recognition technology is becoming an important bio metric solution for people identification in access control as networked access to computer application. Compared to fingerprint, iris is protected from the external environment behind the cornea and the eyelid. No subject to deleterious effects of aging, the small-scale radial features of the iris remain stable and fixed from about one year of age throughout life.

1.2                                          STATEMENT OF THE PROBLEM

In recent years, in line with global trends, the banking sector has faced rising levels of cash card fraud resulting in the subsequent illegal withdrawal of funds from customer accounts. The account-holder is normally held responsible for the loss of funds from their accounts and, as such, the impact of this fraud could be potentially far-reaching.  As a result of this, the banking sector has to embrace bio metrics as the solution to the growing problem of counterfeit ATM cards and ID theft. Among others include

1. Fraudulent card readers, called skimmers are placed over the authentic reader to transfer numbers and codes to nearby thieves.
2. Spy cameras are also used by password voyeurs to collect access codes.
3. In cases of card lost, if the loss is not noticed immediately, consumers may loose all funds in an account.
4. If you forget your pin number, you cannot use the card.
5. The machine can retain your card when the machine malfunctions, when you forget your secret number or if the card is damaged.

1.3                                                   AIM AND OBJECTIVES

The aim of this project work is to simulate an embedded fingerprint authentication system, which is used for ATM security applications. The specific objectives include:

1. To provide a platform that will allow the bankers to collect customers’ finger print.
2. To provide a platform that will allow the bankers to collect customers’ phone number and store them in a centralized database.

III.            To build a system that will forward 4-digit number to the customers’ mobile phone when the finger print reading matches.

1. To provide a platform that allows the customer to run his transaction after the system accepts the code generated.
2. To create a platform that will be able to analyze bio metric data in the global image analysis.

1.4                                                   SCOPE OF THE STUDY

This study is on implementing ATM security using the finger print. There is a centralized database to take care of customers’ personal and bio metric data. The system is designed to query the database by inputting a user finger print and if it matches with the one in a system it will generate a 4-digit number that will enable the user to continue with his transactions.

1.5                                           SIGNIFICANCE OF THE STUDY

The current system of passwords and pin numbers needed to access financial services has drawn a lot of criticism of late due to the increasing incidents of hacking. The system is at the mercy of hackers, who use the hacked data to draw funds from the victims account. This is where Bio metrics with its foolproof system comes in. Some of the reasons for building this system include:

1. Increase security – Provide a convenient and low-cost additional tier of security.
2. Reduce fraud by employing hard-to-forge technologies and materials. For e.g. minimize the opportunity for ATM fraud.
3. Eliminate problems caused by lost ATMs or forgotten passwords by using physiological attributes. For e.g. prevent unauthorized use of lost, stolen or “borrowed” ATM cards.
4. Replace hard-to-remember secret digits which may be shared or observed.

Integrate a wide range of bio metric solutions and technologies, customer applications and databases into a robust and scalable control solution for facility and network access

5. Make it possible, automatically, to know WHO did WHAT, WHERE and WHEN!

1.6                                       ADVANTAGES AND LIMITATIONS

One advantage of  finger-scan  technology  is  accuracy.  Identical matches  are  nearly  impossible  since fingerprints contain a large amount of information making it unlikely that two fingerprints would be identical. Fingerprint technology has another advantage offered by technology; the size of the memory required to store the bio metric template  is  fairly  small.  There  are  some  weaknesses  to  finger-scanning,  most  of  which  can  be mitigated. There is a fraction of the population that is unable to be enrolled. There are certain ethnic groups that have lower quality fingerprints than the general populations. Testing has shown that elderly populations, manual laborers,  and  some  Asian  populations  are  more  difficult  to  be  enrolled  in  some  finger-scanning  systems  [3].
Another problem is that over time, sometimes in as short a period as few months, the fingerprint characteristics of an individual can change, making identification and verification difficult. This problem is seen with manual
workers who  work  extensively  with  their  hands.  There  are  also  privacy  issues  attached  to  finger-scanning technologies.  Some  fear  that  finger-scans  may  be  used  to  track  a  person‟s  activities.  Others  fear  that  data collected  may  be  improperly  used  for  forensic  purposes.  Paying  with  a  government  meal  card  at  checkout instead  of  with  cash  would  identify  the  student  as  a  program  recipient.  The solution  was  for  the  school  to provide students the option of using a finger-scan peripheral to purchase meals [3]. At the end of each month, a bill is sent to their parents for payment or to the free food program for reconciliation.