Biomedical Signal Processing

Sources and types of biomedical signals are presented along with the objective of signal processing. Concepts of random variables, probability functions, and functions of random variables are introduced and expanded to random processes (the "sources" of random signals), moment functions, correlation, convolution and coherence. Parameter estimation based on time-limited random signals along with the property of stationarity and methods for assessment of stationarity are presented.

For assessment of power spectral density, the classical (Fouriere-based) and modern (parametric) methods are studied along with random signal modeling.

Common electrophysiological signals, their properties and common signal processing approaches (EKG, EMG, EEG) are presented. Optimal and adaptive filtering are introduced in the context of noise removal. Cepstrum and homomorphic deconvolution and wavelet transform as a method for time-frequency analysis of nonstationary signals are also covered.

To get insight into principles of random processes in relation to signal processing applications. Understanding of theoretical background of various methods for biomedical signal processing and to recognize practical usefulness of these methods for extraction of information from common electrophysiological and other signals of biomedical origin.

Lectures, individual practical lab work, self study. One part of lab work can be replaced by project work (individual or team assignment). Practical work involves application of methods for signal processing on real signals of biomedical origin (signals from clinical environment or students' own signals recorded during lab assignments from other courses).