27-08-2014, 11:47 AM
IRIS RECOGNITION SYSTEM USING SUPPORT VECTOR MACHINE PROJECT REPORT
IRIS RECOGNITION.pptx (Size: 636.49 KB / Downloads: 7)
INTRODUCTION
Identity of a person by using passwords or cards are not altogether reliable.
Can be Forgotten, Stolen, Disclosable, or Transferable
Reliable personal identification infrastructure required.
Biometric technology, based on physical and behavioral features of human body.
Such as Face, Finger Print, Hand shapes, Iris, Palm print, Keystroke, Signature and Voice.
Each biometric technology has its own advantages and disadvantages based on their usability and security.
Among various traits, IRIS Recognition has attracted a lot of attention.
BIOMETRICS
Simply ‘ replacing passwords ’
Derived from Greek words
--‘ bio ’ (life)
--‘ metrics ’ (to measure)
General term to describe a
characteristics or a process.
Science of ‘ identifying ’ or verifying identity of individual.
Overcome the pitfalls of card and knowledge based systems
Definitions and Terminology
Biometrics: Measurement of physiological/ behavioral attributes.
User: Person who presents biometric data.
Template: Feature vector corresponding to user.
Database: Contains templates of many registered users.
Genuine: A person whose claim is genuine.
Impostor: A person whose claim is fake.
Match Score: Similarity score between ‘test’ and ‘template’ data.
Genuine distribution: distribution of genuine scores.
Impostor distribution: distribution of impostor scores.
Threshold: Score that determines whether to accept or not.
FAR: Fraction of impostors exceeding threshold.
FRR: Fraction of genuine falling below threshold.
Training: Biometric system learns the features of users.
Testing: Test sample is checked against the template.
Verification: Test data is checked against reference data (one-to-one match).
Identification: Test data is checked against all registered data (one-to-many match).
PERFORMANCE MEASURES
False Rejection Rates (FRR) - measure the rate of the system to reject the authorized person.
False Acceptance Rates (FAR) - measure the rates of the system to accept the unauthorized person.
Recognition accuracy -Fraction of correct predictions to the total predictions.
Detection Error Trade-off (DET) curve: Plot of FAR versus FRR for various thresholds.
Score density plot: Plot of genuine and impostor score densities
MOTIVATION
Iris recognition is one of important biometric recognition approach in a human identification.
consists of localization of the iris region and generation of data set of iris images followed by iris pattern recognition.
Located iris is extracted from an eye image.
represented by a data set.
Neural Network (NN) is used for classification of iris patterns
IRIS IMAGE PREPROCESSING
1. Iris localization / segmentation:
- isolate the actual iris region in a digital eye.
- approximated by two circles, one for the iris/sclera boundary and the other interior to the first, for the iris/pupil boundary.
2. Iris normalization:
- to overcome imaging inconsistencies.
- iris region is transformed into rectangular region where the iris texture is analyzed.
3. Enhancement :
- Histogram equalization was used for enhancement of image for getting proper intensity
SUPPORT VECTOR MACHINES
structural risk minimization (minimizing classification error).
- determination of the optimal hyper plane.
- transformation of non-linearly separable classification problem into linearly separable problem