Signal ProcessingThis course will introduce you to techniques in processing of one-dimensional signals, with a particular focus on quasi-periodic signals such as speech and heart signals. Throughout four 4-hour meetings, you will gain experience in processing and analyzing speech and heart signals in time, and in the frequency domains.
The first two sessions are dedicated to speech signals, where students get familiar with the characteristics of speech signals, design a pitch detection algorithm and implement a speech compression system using linear predictive coding (LPC). In the subsequent two meetings, attention shifts to heart signals, specifically focusing on photoplethysmography (PPG) and electrocardiography (ECG). In the third meeting, students will learn to pre-process the signal in order to clean it and improve its quality. They will estimate heart pulse rate from PPG signals, utilizing both pre-recorded examples and self measurements. Finally, in the fourth meeting, participants will leverage ECG signals to estimate heart-beat rates and construct a classifier for diagnosing arrhythmia (heart rhythm disorder).
Led by experienced instructors and supported by designated hardware, this course offers a unique opportunity to practice issues such as filtering in the time and in the frequency domain, noise removal, spectral estimation and auto-correlation, feature selection and performance evaluation. By taking this course you will gain practical skills necessary to analyze and interpret signals in diverse contexts.
