Authentication through gender classification from iris images using support vector machine.

Area of Science:

• Biometrics and Pattern Recognition
• Computer Vision
• Machine Learning for Security

Background:

• Soft biometrics like iris patterns offer secure identification with low error rates.
• Current iris-based gender classification methods suffer from low accuracy and high computational demands.
• Iris recognition is valuable for security, authentication, and validation systems.

Purpose of the Study:

• To develop a robust and computationally efficient gender classification system using iris images for authentication.
• To improve upon existing iris-based gender classification techniques by addressing accuracy and complexity limitations.

Main Methods:

• Feature extraction using Zernike, Legendre invariant moments, and Gradient-oriented histogram from iris images.
• Keycode fusion for attribute categorization and creation of a fused feature vector.
• Gender classification employing a Support Vector Machine (SVM) on the fused feature vector.

Main Results:

• The proposed method achieved a 98% gender classification accuracy on the CVBL dataset.
• Demonstrated significantly lower computational complexity compared to state-of-the-art methods like Local Binary Patterns and Gabor filters.
• Validated the effectiveness of invariant moments and SVM for iris-based gender classification.

Conclusions:

• The developed iris-based gender classification approach offers a highly accurate and efficient solution for biometric authentication.
• The method's low computational complexity makes it suitable for real-time authentication applications.
• This technique enhances biometric security by leveraging unique iris features for gender identification.