Security in biometric data using visual cryptography

Abstract

Biometric privacy is crucial in today’s applications since they use sophisticated and sensitive user data for system authentication. Moreover huge privacy risks are arising, as more powerful and accurate biometric readers are being invented and used. This therefore poses major security risks. These threats can take on the form of database attacks, thus, attempting to acquire and maliciously use digital data from data warehouses. Focusing on metrics related to human characteristics, it is vital to ensure security and safe storage of this data as they are not easily replaced or recreated. Through the use of Visual Cryptographic techniques, this dissertation will explore the best methods that one can use to protect biometric data and ensure safety when transferring between recipients and storing across multiple databases. This project also tries to find a balance between efficiency and data integrity by comparing the results obtained from the application created through third-party facial recognition software. Ultimately shedding some light on different types of decrypted images, exploring the monochromatic and colour spectrum via pixel manipulation and image optimisations. Finally the results show that Visual Cryptography improved the quality on biometric data recognition, whilst still maintaining perfect security and image integrity using simple yet rich UI and primitive colour techniques.