FIT2093 Introduction to Cybersecurity
Assignment 1, Part 2: Iris Biometrics Practical Exercise
Due Date: Sunday 8 May 2021 (end of Week 9), 11:59pm GMT
1 Objective
This individual (not group) practical part of Assignment 1 (worth 7 marks out of the 20 marks for As-
signment 1) is designed to improve your understanding of biometric authentication as a classification
problem. It involves performing computations with dataset of sample Iris Biometrics, using the Open
Source IRIS (OSIRIS) reference system. The concept of iris recognition is described in Daugmen’s
paper (Daugman, 2004), while an overview of the OSIRIS system is described in (Othman, Dorizzi,
& Garcia-Salicetti, 2016) and its usage syntax is decribed in the reference manual (Othman, 2013).
You may read these papers (all available on the Assessment page of Moodle) to gain a better under-
standing of how iris recognition and OSIRIS work before starting the assignment. By completing this
part of the assignment, you should be able to perform segmentation, normalisation and classification
using OSIRIS, and analyse classification results.
2 Setup the OSIRIS Server
To do this assignment you can use your device with the following software installed, as also used in
the lab VMs.
• VirtualBox 4.3.12 (or above).
• VirtualBox 4.3.12 (or above) Oracle VM VirtualBox Extension Pack.
Import the Virtual Machine Download a copy of FIT2093 OSIRIS Server.ova file from the
Assessments page on Moodle. This file contains a Linux Fedora virtual machine (similar to, but
slightly different than Ubunutu used in the lab VMs), in which OSIRIS is installed. Start Vir-
tualBox, on the menu bar select File → Import Appliance to open the dialogue box. Choose the
FIT2093 OSIRIS Server.ova file, then click Import. After a couple of minutes, FIT2093 OSIRIS Server
should be imported and should appear in your VirtualBox Manager window.
Start the Virtual Machine In the VirtualBox Manager window, double click to power on the
OSIRIS Server guest OS that you just imported. Choose second option to login, refer to below pic.
Log in with the default user “fit2093” and password: fit2093. OSIRIS Server Student should
now be ready to use.
1
p.s. If you ever need to become root to change configurations, the root password is toor.
2
3 Getting Started with OSIRIS
You are suggested to read the documentation of OSIRIS v4.1 (Othman, 2013) to better understanding
of this software. All commands used in this assignment are bash commands which should be entered
in a terminal. On the OSIRIS Server Student virtual machine, directories for OSIRIS have been
setup as follow:
• /home/fit2093/Iris Osiris v4.1 contains the source code and sample script to run OSIRIS.
This is the directory where OSIRIS executables are stored.
• /home/fit2093/osiris4.1 contains all files customised for this assignment. This is the direc-
tory you use for the assignment.
The /home/fit2093/osiris4.1/images directory contains some images that OSIRIS can process.
In this assignment we will use 50 images in total. It evaluates 10 individuals and per each of them,
there are 5 images. The names of the images to be used in this exercise are:
• Person 1: 01L0001.jpg, 01L0002.jpg, 01L0003.jpg, 01L0004.jpg, 01L0005.jpg
• Person 2: 02L0001.jpg, 02L0002.jpg, 02L0003.jpg, 02L0004.jpg, 02L0005.jpg
• Person 3: 03L0001.jpg, 03L0002.jpg, 03L0003.jpg, 03L0004.jpg, 03L0005.jpg
• Person 4: 04L0001.jpg, 04L0002.jpg, 04L0003.jpg, 04L0004.jpg, 04L0005.jpg
• Person 5: 05L0001.jpg, ...
• Person 6: 06L0001.jpg, ...
• Person 7: 07L0001.jpg, ...
• Person 8: 08L0001.jpg, ...
• Person 9: 09L0001.jpg, ...
• Person 10: 10L0001.jpg, ...
The name of an image has a meaning. The number before the ‘L’ character is the ID of a person,
and the number after the ‘L’ character is the ID of the image of that particular person. For example,
“01L0001.jpg” in this case, means “Image 0001 of person 01”. Please note that this assignment will
use only these 50 images.
3.1 Segmentation, Normalisation and Encoding
OSIRIS v4.1 can perform segmentation, normalisation and encoding all at once. This software
works by taking a script specifying what images to process, how to process and where to store the
results, and perform actions base on the given script. /home/fit2093/osiris4.1/scripts directory
contains the scripts that you can use for this assignment. You can refer to (Othman, 2013), Section
3, to modify the configuration script and the image list in your /home/fit2093/osiris4.1/scripts
directory to process all 50 images. The script should direct OSIRIS to save the results in
/home/fit2093/osiris4.1/output directory. The sample segmentation script and image list file
are provided below, you can refer to (Othman, 2013) to customise the scripts for your need.
3
Listing 1: Sample configuration.ini
###########################
# What is going to be processed
###########################
Process segmentation = yes
Process normalization = yes
Process encoding = yes
Process matching = no
Use the mask provided by osiris = yes
#############################################
# Text file containing the name of all images to be processed
#############################################
List of images = /home/fit2093/osiris4 .1/ scripts/ImageList.txt
##########################################
# INPUTS : load the input datas from which directories ?
##########################################
Load original images = /home/fit2093/osiris4 .1/ images/
#Load parameters =
#Load masks =
#Load normalized images =
#Load normalized masks =
#Load iris codes =
#####################################
# OUTPUTS : save the results in which directories ?
#####################################
Save segmented images = /home/fit2093/osiris4 .1/ output/
Save contours parameters = /home/fit2093/osiris4 .1/ output/
Save masks of iris = /home/fit2093/osiris4 .1/ output/
Save normalized images = /home/fit2093/osiris4 .1/ output/
Save normalized masks = /home/fit2093/osiris4 .1/ output/
Save iris codes = /home/fit2093/osiris4 .1/ output/
#Save matching scores =
######################
# PROCESSING PARAMETERS
######################
Minimum diameter for pupil = 50
Maximum diameter for pupil = 160
Minimum diameter for iris = 160
Maximum diameter for iris = 280
Width of normalized image = 512
Height of normalized image = 64
Gabor filters = /home/fit2093/Iris_Osiris_v4 .1/ OsirisParam/filters.
4
txt
Application points = /home/fit2093/Iris_Osiris_v4 .1/ OsirisParam/
points.txt
###########
# FILE SUFFIX
###########
Suffix for segmented images = _segm.bmp
Suffix for parameters = _para.txt
Suffix for masks of iris = _mask.bmp
Suffix for normalized images = _imno.bmp
Suffix for normalized masks = _mano.bmp
Suffix for iris codes = _code.bmp
Listing 2: ImageList.txt
01L0001.jpg
01L0002.jpg
01L0003.jpg
01L0004.jpg
01L0005.jpg
02L0001.jpg
02L0002.jpg
02L0003.jpg
02L0004.jpg
02L0005.jpg
03L0001.jpg
...
...
09L0005.jpg
10L0001.jpg
10L0002.jpg
10L0003.jpg
10L0004.jpg
10L0005.jpg
After editing the script, run OSIRIS by entering the following command:
/home/fit2093/Iris_Osiris_v4 .1/ src/osiris