An Efficient Siamese Network Based Multi- Biometric Key Distribution Protocol for Cloud Data Security

Abstract

Storage as a Service is one of the functionality of cloud computing.Storing data in cloud has its own advantages such as high availability, low costs etc., but also disadvantages like improper authentication. Features extracted from single type of biometrics of users are used to serve this purpose. Limitations of using single type of biometric systems include insecure key distribution, high computational time and memory.They are applicable only to structured data with limited data size. To overcome all these disadvantages multiple types of biometrics of userscame into existence. Feature extraction, key generation, encrypting cloud data and key distribution is difficult task for many applications as the number of biometrics considered per user increased. To overcome all these limitations an efficient Siamese Network based Multi-Biometric (SNMB) model is proposed Experimental results are tested on different biometric images and the performance of the proposed SNMB model has better computation time(~6%) and less memory(4%) than the conventional cloud security models.