Related Experiment Video
Updated: Jul 3, 2026
11:04
Biomolecular Detection employing the Interferometric Reflectance Imaging Sensor (IRIS)
Published on: May 3, 2011
Improving iris recognition performance using segmentation, quality enhancement, match score fusion, and indexing
Mayank Vatsa1, Richa Singh, Afzel Noore
1Lane Department of Computer Science and Electrical Engineering, West Virginia University, Morgantown, WV 26506-6109, USA. mayankv@csee.wvu.edu
Summary
This study introduces advanced algorithms for iris recognition, enhancing accuracy and speed. The methods improve iris image quality, feature extraction, and matching for more reliable biometric identification.
Area of Science:
- Biometrics
- Computer Vision
- Pattern Recognition
Background:
- Iris recognition systems require robust segmentation and quality enhancement for accuracy.
- Existing methods face challenges with non-ideal iris images and efficient identification.
Purpose of the Study:
- To develop novel algorithms for iris segmentation, quality enhancement, match score fusion, and indexing.
- To improve both the accuracy and speed of iris recognition systems.
Main Methods:
- Curve evolution for iris segmentation using modified Mumford-Shah functional.
- Support-vector-machine-based selection of enhanced regions for high-quality iris images.
- Feature extraction using 1-D log polar Gabor transform (textural) and Euler numbers (topological).
- Intelligent fusion of textural and topological matching scores.
- Indexing algorithm for fast and accurate identification.
Main Results:
- Proposed algorithms demonstrated improved iris recognition performance.
- Reduced false rejection rates through intelligent match score fusion.
- Achieved faster and more accurate iris identification via the indexing algorithm.
- Validation conducted on CASIA V3, ICE 2005, and UBIRIS databases.
Conclusions:
- The developed algorithms offer significant improvements in iris recognition accuracy and speed.
- The combined approach of enhanced image quality, novel feature extraction, and intelligent fusion is effective.
- The system provides a robust solution for biometric identification challenges.