Topic: BusinessManagement

Last updated: March 17, 2019

Project I Report

Aslam Shah (140420111025)
Muaviya Lala (150423111003)
Karnav Parmar (150423111005)
Hushain Rana (150423111008)

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now

Department of Electronics and Communication Engineering

Prof. Chintan Panchal

Sarvajanik College of Engineering & Technology
Dr. R.K. Desai Road, Athwalines, Surat-395001


This is to certify the Project Report entitled “FINGERPRINT BASED PERSON IDENTIFICATION
USING RASPBERRY PI” is prepared & presented by:

Aslam Shah (140420111025)
Muaviya Lala (150423111003)
Karnav Parmar (150423111005)
Hushain Rana (150423111008)

Of B.E. IV Sem VII Electronics & Communication Engineering department during year 2017 -18.

Prof. Chintan Panchal Prof. Maulin Joshi
Guide HOD

Signature Signature

Jury Name: Signature


At first, we would like to express our project guide Prof. Chintan Panchal, electronics and communication engineering scet surat for his constant guidance and introducing the Biometric system to us and also inspiring us. We are also very much thankful to our faculties for their constant support throughout the project work and their help in the clear understanding of ‘Biometric system’.
And we extend our sincere thanks to all of our friends, classmates and some special friends whom I would like to thanks for helping us with the algorithm and moral support in this project.


Aslam Shah (140420111025)
Muaviya Lala (150423111003)
Karnav Parmar (150423111005)
Hushain Rana (150423111008)


In our B.E final year project we focus on the current techniques for the fingerprint recognition. Every Human fingerprint has some certain details on it. We said it as minutiae, which can be used as a unique identity of a person if we recognize it in a well manner. The main aim of this project is to design a complete system for person identification from its fingerprint with the help of raspberry pi using minutiae matching technique. So for having a good quality minutiae extraction the fingerprint image is first pre-processed by image enhancement, Fast Fourier Transform and binarisation and then segmentation is done to differentiate certain area of the fingerprint followed by minutiae extraction which includes ridge thinning and minutiae marking after that we have to do a post- processing operation which includes removal of false minutiae, H-breaks, and isolated points.
Now, we need to do final process that is ‘minutiae matching’, in minutiae matching we match the processed fingerprint image with the image which is stored at database. For all these we develop matching algorithm which is for minutiae matching. This algorithm has ability to find the difference between input minutiae and the stored minutiae pattern .We can then check the performance of the system by taking different fingerprints of different persons.


1.1 What is fingerprint? 1
1.2 Motivation and challenges 3
1.3 Using biometrics 4
1.4 Why we use fingerprints? 4
1.5 What do you mean by Fingerprint Recognition 5
1.6 Approaches for Fingerprint Recognition………………………………………….6
2.1 System level Design 8
2.2 Algorithm Level Design………………………………………………………….9
3.1 How does fingerprint recognition work?………………… ……………………..11
3.2 Fingerprint identification system Flowchart………………………………………11

4.1 Requirement………………………………………………………………………..14
4.2 Step for installation of OS on R- PI……………………………………………… 16

4.3 Approaches for fingerprint recognition…………………………………………… 17




1.1 Finger-print

A finger-print is a pattern of feature of a finger as shown in the figure 1.1.1 given below. It is believed that each different fingerprint of different person in this world is unique with permanent unique characteristics as per the strong evidence. That’s why nowadays fingerprints are used for various identification and investigation process. In today’s era we uses fingerprints for many purposes like, fingerprint based attendance system, fingerprint based verification etc.

Figure1.1.1 A fingerprint image using an Optical Sensor
A fingerprint has many ridges and furrows. We can see good similarities between these ridges and furrows for a taken small local window, like average width and parallelism.
However, on the basis of intensive research on fingerprint recognition, we come to the conclusion that fingerprint are not recognized using their ridges and furrows, but minutiae plays a vital role over here, which are characterized by some abnormal points on the ridges as shown in below figure (1.2.2). and in figure (1.2.3 and 1.2.4) we can see a variety of minutiae. Although we can have a variety of minutiae types as per with literature, but two types of minutiae are mainly used and most significant and we?ll also extend these in our project. In which, one is called „termination? which can be characterized as the immediate ending of a ridge and the other one is called „bifurcation? which can be characterized as the point on the ridges where two branches are bifurcated as shown in figure (1.1.2)

Fig 1.1.2 : Minutiae (ridge termination and bifurcation)

Fig 1.2.3: Minutiae points on a fingerprint
Fig 1.2.4 Different types of minutiae

1.2 Motivation and Challenges:

What we see in every organization, whether it is a business organization, educational colleges or institution, it has to classify each individual on the entrance gate either for attendance or for their department identification for an effective functioning in a well- mannered way. So, designing and automatic system which can verify their identity just by recognizing their fingerprints will reduce the man-work and also introduce a time saving mechanism with better operating capabilities we can even made it a lot faster and get some personal information of the person also using online databases. It’s our responsibility to make an indigenous design which can operate with such constraints with an improved fingerprint recognition system.

1.3 Using Biometrics:

Some of the instruments used for unique identification of humans include biometrics identification systems known mainly for identification and verification. Here, we will use biometrics as for access management and access control of identity. So, use of biometrics in fingerprint recognition is secure and easy for image acquisition step in fingerprint recognition as well. We can see a lot varieties of biometric systems like face detection, fingerprint recognition, iris recognition, voice recognition, palm recognition etc. in our project we will only go for fingerprint recognition using a biometric device to capture fingerprint image.


Fingerprints are considered as a unique identification of a person and due to easy access it?s the best and one of the fastest method used in biometric identification systems. They are unique, so secure and reliable to use and doesn?t change for one in a lifetime. And beside these things fingerprint recognition specially using minutiae matching technique is cheap, reliable and accurate up to a satisfactory limits.
Hence, fingerprint recognition is being widely used in both civilian and forensic applications. If, we will compare with other biometric devices then fingerprint recognition devices will hold the maximum market share and are most proven ones also. And we can also say that it?s not only faster than other biometric devices but it?s energy efficient also, as it consumes very less energy.
Here we are going with „minutiae matching technique? for fingerprint recognition which can be divided into two sub-domains: one can be classified as fingerprint verification and other one can be classified as fingerprint identification (fig 1.5.1). And as we are using a different technique from manual approach for fingerprint recognition, we can say this technique system as „Automatic Fingerprint Recognition System? (AFRS) , which is coded using Matlab(matrix-laboratory)

FIG 1.5.1 Verification and Identification

Fingerprint verification is the final step of fingerprint recognition used to verify the identity or say authenticity of one person by his fingerprint. In this mechanism what we are going to do is to have a user who will give his fingerprint along with his unique identification number. Now, the system will search his fingerprint using the unique
identification number an d if the fingerprint matches then result will be positive otherwise negative. Basically, it is based on design principle of AFAS 39999(Automatic Fingerprint Authentication System)
Fingerprint Identification deals with the specifying the identity of a person by his fingerprint without knowledge of the identity of the person. In this Method generally a large database is stored and the fingerprint taken by the user is matched with the whole database fingerprints. Its? uses can be seen in criminal investigation cases and it is based in the design principle of AFIS (Automatic Fingerprint Authentication System).
But, in the end we can say that all fingerprint problems either verification or identification all start with the same technique of fingerprint recognition and are based on a well-defined representation of a fingerprint. As for the evidences as far as the fingerprints are unique , either we use 1 to 1 verification or we use 1 to m identification case, both will start with the same procedure with some straightforward and easy steps


There are two types of representation, we can classify for fingerprints which make the two approaches for fingerprint recognition.
The first approach is „Minutiae based?, which represents the fingerprint by its local feature mainly stated are the two minutiae features as termination and bifurcation. This approach has been studied intensively and is mainly followed in current fingerprint recognition instruments. My project is also related to this technique with possible corrective measures based on experiments.The second approach uses image-based method; it basically tries to match the whole fingerprint image using the global features. It is the latest and advanced techniques for fingerprint recognition and many researches are still going on to convert it into a cheaper and easy method of use. So, we can say that it?s an emerging technique. My project does not include this approach and so no further studies will be seen on this approach in my thesis.


A fingerprint recognition device is constituted of a fingerprint acquiring device for image acquisition step, minutia extractor for extraction of valuable minutiae and minutiae matcher for matching the minutiae as shown in FIG. 2.1.1

FIG 2.1.1: Basic Model of Fingerprint Recognition system

For fingerprint acquisition step the acquiring devices which are optical or semiconductor sensors, are widely used. They exhibit high accuracy and efficiency unless and until user?s finger is too dirty or dry. However, we are using an online database for the result verification in in our project. We will go through the minutiae extractor and minutiae matcher modules in the next part where algorithm level design is explained and we will extend the discussion in other subsequent sections as well.


A minutiae extractor can be implemented as a three stage approach and is widely used by researchers. They are preprocessing, minutiae extraction and post processing stages

` FIG 2.2.1 Minutiae Extractor

For the image preprocessing steps, we have used histogram equalization followed by Fast Fourier Transform to do the image enhancement and then image binerization is done by locally adaptive threshold method. The image segmentation has two parts, one is ridge flow estimation and other one is by the extraction of region of interest(ROI)using morphological methods. Most of the pre-processing stages used here are a part of standard studies taken by many researchers but here they are carried in our project on basis of a lot of practical results taken by us.
For minutiae extraction stage, we take the help of a three thinning algorithm and we got a morphological thinning operation with a very fine thinning quality and high efficiency. Then the minutiae marking is a simple one just some regular MATLAB functions can handle them.
For the post-processing stages, a better and a very fine algorithm is required to remove false minutiae like H-breaks and isolated points etc.

The basic concept for minutiae matcher is to take a reference point or line then decide the origin for the co-ordinates and now translate and rotate the whole image in order to get the match. So, it first takes any two random minutiae as a reference pair and then matches their associated ridges. If, the ridges are matched very well then both the fingerprints are aligned and matching is done for all the extracted minutiae (FIG 2.2.2)



An identification system can be defined as the one which helps in identifying the individual from many people available. It generally involves matching available biometrics feature like fingerprint with the fingerprints which are already enrolled in the database.

In real life, the fingerprint images that are captured are not of optimum quality. So, we need to enhance their quality and remove noises. We also extract features likes minutiae and other details for matching. And if minutiae sets are matched with those in the database, we say it an identified fingerprint. And after matching we go for post-matching steps which includes retrieving details of the user from the online store and show it. A flowchart is shown in the next section.

A complete methodology of our fingerprint identification system is briefly shown here in a flowchart. Each of whose steps are explained in later chapters FIG 3.2.1.


Block Diagram



? Raspberry Pi
? Fingerprint sensor (optional)
? Monitor/keyboard/Mouse


? Rasbian OS
? Python


? Rasbian OS installation
? Python basic codes


4.2 Step for installation of OS on R-PI:

1. Download os NOOBS/RASBIAN
2. FORMAT SD card
3. Write img file on sd card (win32diskimager)
4. Insert sd card in pi
5. Connect with monitor
6. START the pi using adaptor
7. It will automatically boot the os
8. After booting process desktop of is displayed on PI


There are two types of approaches for fingerprint recognition.
The first approach is “Minutiae based”, which describe the fingerprint by its local feature like termination and bifurcation etc. This approach has been studied more deeply is carried out in our current fingerprint recognition system. The other approach uses method which is image-based; it normally tries to match the whole fingerprint image using the global features. It is the latest techniques for fingerprint recognition nowadays but the problem is that it is high cost system. We are not using this in our project and so no further discussion.


I'm Piter!

Would you like to get a custom essay? How about receiving a customized one?

Check it out