Artifacts processing for sleep stage classification

When a patient falls asleep, he transits by different sleep stages, which characterize the quality of his night. To determine these sleep stages, the technicians visually analyzes the polysomnographic signals (PSG) witch have different aspects according to these stages. However, this task requires a lot of time. This is why one generally tries to class them automatically. Then, one is generally faced with the problem of artifacts witch often interfere during feature extraction. Their detection and processing are therefore mandatory before any other classification operation.