Context-aware Multi-lead ECG Compression Based on Standard Image Codecs

The use of telemedicine capabilities to manage aged and cardiac chronically ill patients is going to become a common practice. Usefulness and diagnostic value of classical ECG monitoring and recording can be enhanced by jointly collecting and analysing data detected by other sensors (e.g. movement detectors) which enable to associate specific cardiac events with the patient's environment and activity at the time epoch the cardiac event appears. In this scenario, characterized by a continuous growth of data volume to be stored and transmitted, data compression plays a crucial role. In this paper we propose a compression method aimed at preserving and exploiting the different diagnostic importance of different ECG segments, making smart use of context information, i.e. information about the patient's condition. Specifically, we focus on a 2D compression method that exploits the features of JPEG2000 compression and we propose a novel paradigm for context-adaptive compression of ECG data.