Biometric Signature Authentication Scheme with RNN (BIOSIG_RNN) Machine Learning Approach

Abstract

In healthcare cloud base management is consider as effective way for data management in health care. Health care management underlies major problem of data security. Data security increases fraudulence activity in medical, tax, and bank, insurance. Data retrieval improves data security for secure data access in medical and health care service. Hence, it is necessary to enhance the security in health care data base for password and token theft. Here, cloud based healthcare data management service, namely, HealthCloud is proposed that ensures higher level of security for e-medical data through biometric behavioral signature authentication. A novel Biometric Signature Authentication scheme is being proposed using Recurrent Neural Network (BIOsig_RNN) for secure and retrieval of data access. For collected biometric signature samples BIOsig_RNN utilizes Machine Learning framework is utilized to train the signature samples for better authentication. It uses recurrent neural network (RNN) to support the Machine Learning framework based on statistical dataset learning. Experimental analysis of proposed approach exhibits increased sensitivity and specificity rate of 0.98 and 0.95, respectively. Comparison with other state of art methods shows that the HealthCloud management system attains better performance as compared to existing methods and system.