Multibiometric System for Iris Recognition Based Convolutional Neural Network and Transfer Learning

Abstract

Multimodal biometric methods have been commonly used by several implementations because of its capability to work with a variety of important drawbacks in unimodal biometric methods, such as noise affectability, populace coverage, intraclass variety, vulnerability to spoofing, and non-universality. In this research, a multimodal biometric realtime method is suggested depending upon the design of a deep learning model for pictures of a person’s (right & left) irises. This system has been implemented by combining the characteristics of convolution neural networks and transfer learning techniques. Through this research, the training system focused on a collection of the back_propagation technique with Adam’s optimization approach utilized to modify weights and adjust learning rates during the learning process. The efficiency of the system is examined on two public datasets obtained in various conditions: IITD and CASIA-Iris-V3 Interval. The implemented system gives an accuracy of 99% for both left & right IITD iris datasets and accuracy of (94% and 93%) for the left and right iris for CASIA-iris-V3 interval datasets respectively after training. An OpenCV library for image pre-processing, Keras, and sci-kit learn python libraries for feature extraction and recognition has been utilized.

Monitoring Student Attendance Through Vision Transformer-based Iris Recognition

“…In [18], the authors presented a multimodal biometric system utilizing Convolutional Neural Networks (CNN) and transfer learning for iris recognition. It aims to overcome limitations in unimodal biometric methods by focusing on deep learning models for analyzing both left and right irises.…”

Detection and Classification of The Osteoarthritis in Knee Joint Using Transfer Learning with Convolutional Neural Networks (CNNs)

“…(Visual Geometry Group) is one of the popular CNN models previously trained on a large dataset from ImageNet. K. Simonyan and A. Zisserman created this model in 2014 [25,26]. The input image to the VGG model is a 224 X 224 RGB image that is fixed in size.…”

Personal Authentication System Based on Iris Recognition and Digital Signature Technology

“…Image Acquisition: This is an important stage in this system since all the next stages highly depend on the output of this stage. At first, the iris dataset was utilized to test the system [1]. After several attempts, images can be directly taken from the camera and entered the port of Raspberry Pi 4 Model B successfully, offering a complete authentication system.…”

“…As the biometric identification process offers several advantages over conventional identification techniques, it has become widely accepted as the method for the unique identification of the user. It becomes critically important for the biometric identification process to be safe from attacks in applications where security is highly important, such as ecommerce [1]. Identity theft has been a major concern in the internet age, and uniquely identifying an individual is proven to be their best defense.…”