Selective encryption on ECG data in body sensor network based on supervised machine learning

Body Sensor Networks (BSNs) are developing rapidly in recent years as it combines the Internet-of-Things (IoT) and data analytic techniques for building a remote healthcare system. However, as BSNs are implemented on the existing wireless communication systems, the security and privacy in the BSN are facing many challenges. Performing standard encryption schemes on the health data before outsourcing at the sensors' ends are not suitable for this BSN environment as it is costly both in energy and time consumption for the BSN sensors. Traditional lightweight encryption schemes such as Selective Encryption (SE) schemes could be used in this environment by reducing the data volume to be encrypted. In this paper, we re-define the SE schemes in a practical scenario of securely outsourcing the electrocardiogram (ECG) data in the untrusted BSN environment. Specifically, if the ECG data is outsourced for disease classification based on a machine learning model, we prove that the classic SE schemes are not the correct designs. Then, we give our SE design based on this classification use case to protect the ECG data against illegal classification at the attacker sides which further protects the patients' data privacy. Intensive tests are experimented to prove the effectiveness of our proposed SE method.