Location

https://ccse.kennesaw.edu/computing-showcase/cday-programs/spring2021program.php

Streaming Media

Document Type

Event

Start Date

26-4-2021 5:00 PM

Description

This work investigates different iris normalization techniques to compare their performance including elliptical normalization and circular normalization after frontal projection of off-angle iris recognition. Elliptical normalization samples the iris texture using elliptical segmentation parameters. For circular unwrapping, we first estimate the gaze deviation using ellipse parameters and the image will be projected back to frontal view using perspective transformation. Then, we segment the transformed image and normalize using circular parameters. We further investigate if: (i) elliptical normalization or circular unwrapping recognition performance is higher, and (ii) the two segmentations methods in circular unwrapping increase the recognition efficiency. Based on the preliminary results, the elliptical normalization method shows slightly better recognition performance in off-angle iris images with 2.2% decrement in the equal error rate. The motivation of this research is to provide guidance in the construction of a recognition framework for off-angle iris images with the analysis and comparison of different normalization methods. MATLAB was used to calculate the hamming distance and accuracy of each normalization method and to construct the plotted graphs for display. Based on the preliminary results, the elliptical normalization method shows slightly better recognition performance in off-angle iris images with 2.2% decrement in the equal error rate. In addition, perspective projection shifted the distributions of intra and inter class Hamming distances to left where its average intra-class and inter-class Hamming distance are 0.3070 and 0.4891, respectively compared with 0.3082 and 0.4900 for elliptical normalization.Advisors(s): Dr. Mahmut Karakaya mkarakay@kennesaw.eduTopic(s): Security

Share

COinS
 
Apr 26th, 5:00 PM

GR-1 Compare Two Off Angle Normalization

https://ccse.kennesaw.edu/computing-showcase/cday-programs/spring2021program.php

This work investigates different iris normalization techniques to compare their performance including elliptical normalization and circular normalization after frontal projection of off-angle iris recognition. Elliptical normalization samples the iris texture using elliptical segmentation parameters. For circular unwrapping, we first estimate the gaze deviation using ellipse parameters and the image will be projected back to frontal view using perspective transformation. Then, we segment the transformed image and normalize using circular parameters. We further investigate if: (i) elliptical normalization or circular unwrapping recognition performance is higher, and (ii) the two segmentations methods in circular unwrapping increase the recognition efficiency. Based on the preliminary results, the elliptical normalization method shows slightly better recognition performance in off-angle iris images with 2.2% decrement in the equal error rate. The motivation of this research is to provide guidance in the construction of a recognition framework for off-angle iris images with the analysis and comparison of different normalization methods. MATLAB was used to calculate the hamming distance and accuracy of each normalization method and to construct the plotted graphs for display. Based on the preliminary results, the elliptical normalization method shows slightly better recognition performance in off-angle iris images with 2.2% decrement in the equal error rate. In addition, perspective projection shifted the distributions of intra and inter class Hamming distances to left where its average intra-class and inter-class Hamming distance are 0.3070 and 0.4891, respectively compared with 0.3082 and 0.4900 for elliptical normalization.Advisors(s): Dr. Mahmut Karakaya mkarakay@kennesaw.eduTopic(s): Security