Machine Learning Based Approach for Efficient Multibiometric Indexing

Biometric identification systems have become essential for secure and efficient authentication methods. This project focuses on creating a streamlined system for accurate, rapid, and space-efficient identification by combining iris and fingerprint biometrics. In this approach, iris and fingerprint images are fused using mean and standard deviation values to produce a unified biometric representation. Support Vector Machine (SVM) and Principal Component Analysis (PCA) algorithms are applied to classify and reduce the dimensionality of the fused data, enhancing the efficiency and accuracy of identification. Unique hash codes are generated from the processed data, allowing for fast and storage-efficient search within the database. By verifying these hash codes against stored profiles, the system determines if an individual is authorized, achieving an accuracy of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{8 0 \%}$</tex>. This approach, combining SVM and PCA, enables a highly secure, scalable, and effective multibiometric indexing system, ideal for applications where speed and reliability are critical.