Secure Personal Identification through Fundus Image-Based Transfer Learning

Biometric identification systems are very important in security. It is easy for an intruder to collect someone's fingerprints, facial features, iris patterns, or voice recognition and then inject it into a biometric identification system. It points out a potential vulnerability in traditional biometric methods. The purpose of this study is to use fundus images to create secure and reliable digital templates for personal identification. We use transfer learning to identify individuals based on fundus images. Transfer learning involves leveraging knowledge gained from one task to improve performance on a related. It enhance the efficiency and effectiveness of machine learning models. We use RIDB (Retinal Identification Data Base), which are publicly available databases. The proposed scheme provides a higher accuracy rate of 99.998% for RIDB. This suggests that the developed model is highly effective in accurately identifying individuals based on their retinal characteristics, presenting its potential as a secure and reliable method for personal identification.