Electrocardiogram signals (ECGs) measure the electrical activity of the heart. These time series are used by physicians to monitor patients and to diagnose various cardiac diseases. In particular, the duration of certain patterns and the time interval between specific events are intensively used. Because ECGs may last for hours and contain thousands of beats, automated tools exist to help physicians to segment ECGs.
I have published several works about automated segmentation of ECGs with hidden Markov models (HMMs) and wavelet transforms. In particular, I proposed methods to improve transition modelling and to make HMM inference robust to label noise.
- Frénay, B., de Lannoy, G., Verleysen, M. Label Noise-Tolerant Hidden Markov Models for Segmentation: Application to ECGs. In Proc. ECML-PKDD 2011, p. 455-470.
- Frénay, B., de Lannoy, G., Verleysen M. Improving the transition modelling in hidden markov models for ECG segmentation. In Proc. ESANN 2009.
- Frénay, B., de Lannoy, G., Verleysen M. Emission modelling for supervised ECG segmentation using finite differences. In Proc. MBEC 2009, 1212–1216.
- de Lannoy, G., Frénay, B., Verleysen M. Supervised ECG delineation using the wavelet transform and hidden markov models. Proc. MBEC 2009, 22–25.