Secure Multi-Modal Biometrics With Cnn: Integrating Signature, Face, and Fingerprint Authentication
Irakam Chaitanya, Olla Yaswanth Charan, G. Keerthiga · 2025
This work outlines a card-less ATM security system based on a biometric authentication of fingerprint, face and signature recognition based on deep learning algorithms developed in MATLAB. The system employs CNNs for identification that means steady biometric input from a single person to let in the system. This approach goes further by utilizing many biometric modes thus improving on security and at the same time avoids pitfalls that are associated with the use of ATM cards and numerical PINs that are very easy to hack. The developed system is focused on reliability and clear user interface which ultimately leads to an improved, more convenient transaction flow. The findings of this work create the basis for future developments in ATM security as well as also increasing the general confidence of users in biometric solutions for the completion of financial transactions and forms a basis for expansion of safe banking.