11-04-2011, 10:43 AM
PRESENTED BY:
Prathusha Prakash
[attachment=11984]
Introduction
Iris recognition technology is used to identify individuals by photographing the iris of their eye.
It falls under a category of technology known as biometric-based authentication, also called biometric security.
Iris recognition technology works by combining computer vision, pattern recognition, and optics.
Biometrics
Uniquely identify people according to physical or behavioral traits
Types of Biometric systems
Iris
Speech
Face
Finger print
Human Eye
Iris
Colored ring
Uniqueness and randomness in its patterns
Why Iris Authentication?
Externally Visible
Unique to all individuals
Protected by cornea and hence non-invasive
Stable
How Iris Authentication makes a better option?
Iris recognition technology has become popular in security applications because of its ease of use, accuracy, and safety. Its most common use is controlling access to high-security areas.
Use of the iris in authentication is one of the best ways of meeting high risk situations.
When the iris scan is done in the presence of an enterprise employee, the chance that the person is who they claim to be is very high.
Iris recognition technology can confirm someone's identity within a few seconds.
Flow chart proposed for Iris Authentication
Explanation
First, a black-and-white video camera zooms in on the iris and records a sharp image of it.
A frame from this video is then digitized.
Iris recognition technology is capable of recording this image from as much as 16 inches (40.64 centimeters) away, so no physical contact is necessary
Glasses or contact lenses do not interfere with the operation of iris recognition technology
Iris Localization
Detect the Inner Boundary of Iris
Multi-scale edge detection is used to find the edge of pupil.
Relate edges to local maxima.
Delete the short line and single points.
Finally, Hough transform locates boundary of the pupil labeled by white line.
Detection of outer boundary
Edge detection and deletion of short line is carried out.
Most iris radius is between 80 and 150.
Finally, Hough transform is used to locate the outer boundary of iris.
Iris Normalization
The iris image is transformed from polar coordinates to a 512×80 fixed size rectangular image to reduce the effect of iris dilation and contraction, of which 512×48 will be coded.
The non-uniform background illumination is finally homogenized
ROI Segmentation
Only two regions of interest (ROIs) that between−35° ~ +10° and +175° ~ +215° of iris ring are used forauthentication.
Feature Extraction
Gabor Filters
To obtain filtered image
Feature Selection with GA
Extracting useful features for optimal solution
Feature Classification with SVM
Classify the data accordingly thus helping in identification and verification
Gabor Filters
A Gabor filter is a linear filter used in image processing for edge detection. Frequency and orientation representations of Gabor filter are similar to those of human visual system, and it has been found to be particularly appropriate for texture representation and discrimination.
A `Gabor filter` is a linear filter whose impulse response is defined by a harmonic function multiplied by a Gaussian function.
h(x,y)=s(x,y)g(x,y)
Before feature extraction the two ROI segmented previously are combined into one ROI
Then divided into two images: up part and down part.
From the inner boundary to outer boundary, the texture information of iris changes from fine-to-coarse.
So, a small scale Gabor filter is used to capture the texture of up part, and relative big scale Gabor filter is used to capture the texture of down part.
Genetic Algorithm
Inspired by Darwin's theory of evolution.
Genetic Algorithms are a search method that can be used for both solving problems and modelling evolutionary systems.
It is Heuristic-just estimates the solution.
Works on a population of possible solutions, which other heuristic methods use a single solution.
Why Genetic Algorithm?
Can quickly scan a vast solution set.
Bad proposals do not effect the end solution negatively.
Inductive nature of the GA-it has its own internal rules.
Useful for complex or loosely defined problems.
Support Vector machines
Support vector machines (SVMs) are a set of related supervised learning methods used for classification and regression.
A support vector machine constructs a hyper plane or set of hyper planes in a high or infinite dimensional space, which can be used for classification, regression or other tasks.
Areas Of Application
Airport
Immigration Office
Banks
Forensics
Organization
Secure Financial Transactions
Personal Certificates
UAE immigration Inspection
Control aid distribution for Afghan Refugees
Netherlands Passport free immigration
Experimental Results
Small image database including 140 iris images from 20 users as testing sample
4 iris images of an individual as training sample
Iris Localization
Parameters is specific,hence method can segment iris effectively.
Feature Selection
1 if feature of related block is selected
0 if feature of related block is not selected
Correct Verification Rate(CVR) can be calculated using False Rejection Rate(FRR) and False Acceptation Rate(FAR) which indicates the performance measurement of this method.
CVR = 1 – (FAR + FRR)
For proving the superiority of the proposed method, we compare the CVR between feature selection and without feature selection in our small sample database.
Comparison Results
Advantages
Patterns apparently stable throughout life.
Accuracy rate is higher compared to other techniques.
Iris textures are completely unique. Even identical twins have different iris textures.
An internal organ such as the iris is well protected and is not susceptible to wear as fingerprints are.
Surgical procedures do not change the texture of the iris, which is used as the distinguishing factor by iris recognition systems.
Disadvantages
Iris scanning is a relatively new technology and is incompatible with substantial investment.
Iris recognition is very difficult to perform at a distance larger than a few meters.
As with other photographic biometric technologies, iris recognition is susceptible to poor image quality.
Iris-recognition technology might help governments to track individuals beyond their will.
Conclusion
Novel Iris localization approach and user specific feature selection scheme has been used.
Each user establishes specific parameters.
Experimental results show that proposed method can achieve lower error rates.
Thus proving its effectiveness and feasibility.
It is highly accurate as well..