ECG Signal Compression Using Adaptive Hermite Functions

In modern medical science evaluation of electrocardiogram (ECG) has proven to be an important task for doctors. These signals contain valuable information on the patients’ condition; however analysis of them has encountered numerous challenges, such as storage of long-term recordings, filtering, and segmentation of signals. Resolving these problems is important to ensure a high quality diagnosis. In this paper we propose an ECG analysis method which provides adequate solutions to all of these challenges. The proposed method is based upon the approximation theory in Hilbert spaces. Namely, using the affine transforms of orthonormal Hermite systems, the approach optimizes two free parameters. This is done in order to achieve the best approximation of the ECG signal using a fixed number of Fourier coefficients. The process of optimization is done using Particle Swarm Optimization (PSO), Nelder- Mead (NM) simplex method, and Monte Carlo (MC) algorithm which are embedded into a matching pursuit framework. The former procedure guarantees both good compression ratio and high accuracy, while the latter segments the heartbeats. As it is shown by experiments, the proposed method achieves better results than previously known approaches.