A method for recognizing noises in pulse waves

A method is proposed for recognizing noises in pulse waves (PWs), which play an essential role in clinical pulse diagnosis. PWs have also been used in traditional Chinese medicine for thousands of years. We follow the principle that most noises appear as pairs of adjacent sampling points. These pairs satisfy five certain criteria. The principle and the criteria are based on observation and statistical analysis. Our proposed method has a noise recognition ratio of 82% and can help a lot for other physiological signals.