Adaptive computing-based biometric security for intelligent medical applications

Abstract

For intelligent medical systems, clinical information security is one of critical requirements. Recently, random binary sequences (RBSs) based on interpulse intervals (IPIs) were applied as secret keys to secure medical information in medical applications. Most of the existing RBS generation methods acquire a uniform quantity of bits per IPI, thereby requiring more processing time. However, the functional capacity of humans influences their heart rate variability, thereby resulting in the extraction of adaptive entropic bits per IPI across individuals. Therefore, adaptive computing-based RBS generation method is proposed to extract a variable number of bits on the basis of the heart rate (HR) bands of individuals to provide a balance between processing time and security in WBSNs. The proposed method is evaluated by using ECG recordings of 126 subjects with dynamic scenarios. Our experimental results that 128-bit RBSs generated by applying the proposed method can be used as secret keys for entity identifiers or patient’s data encryption for securing intelligent medical applications. In this study, the hamming distance metric is used to measure the uniqueness of the generated RBSs, and randomness of RBSs is computed by means of statistical tests, for different HR bands. Furthermore, the processing time of the proposed method for RBS generation shows improvement compared with the conventional techniques. The proposed approach is approximately three times faster for 55 ≤ HR < 80 and approximately two times faster for 80 ≤ HR < 105 and 80 ≤ HR < 105 than the existing IPI-based RBS generation techniques. Therefore, this study has got real-time significance for smart healthcare applications.