Filtering for Removal of Artifacts

This chapter introduces the nature of the artifacts that are commonly encountered in biomedical signals. Several illustrations of signals corrupted by various types of artifacts are provided. Details of the design of filters, spanning a broad range of approaches, from linear time-domain and frequency-domain fixed filters to the optimal Wiener filter to adaptive filters, are also described. The chapter also investigates problems posed by artifact, noise, and interference of various forms in the acquisition and analysis of several biomedical signals. Random noise, structured interference, and physiological interference have been identified and analyzed separately. Attention has been drawn to the different characteristics of various types of noise, such as frequency content and nonstationarity. It is important to observe that each practical problem needs to be studied carefully to determine the type and characteristics of the artifact present. It is unlikely that a single filter will address all of the problems and the requirements in a wide variety of practical situations and applications. Irrespective of one's expertise in filters, it should be remembered that prevention is better than cure: most filters, while removing an artifact, may introduce another.