Bioacoustic Signals

AbstractBiological sounds, such as lung sounds, heart sounds, bowel sounds, and joint sounds, emitted from the human body during its functioning provide an easy and noninvasive way of indicating existing pathology. By applying advanced signal processing techniques to the so‐called bioacoustic signals (BAS), useful information of high diagnostic value can be revealed. The purpose of this review is an in depth exploration of the properties of BAS, starting from their categorization, their characteristics and relation to human pathologies, and moving on to the recording standards that should be followed for their accurate acquisition, and focusing on the most current BAS analysis techniques. The latter refer to important signal processing issues, such as de‐noising, detection, modeling, feature extraction, and classification, involving a variety of methodologies, such as spectral analysis, wavelet transform, higher‐order statistics/spectrum, lower‐order statistics, fractal dimension, higher‐order crossings, reduced interference time‐frequency distributions. As a result, the way important problems in BAS analysis are approached with modern mathematical techniques is presented. Some descriptive examples along with an updated bibliography are provided to the reader as a means for thorough exploitation of BAS, because they cover advanced applications in the field. Consequently, the current pathways and trends in BAS analysis are introduced, giving rise to their objective evaluation and interpretation. However, they reveal the call for a common interest in interdisciplinary undertakings, both by biomedical engineers and by clinicians. Although it is unknown when traditional clinicians accept alternative approaches in everyday clinical practice, the results shown here are the outcome of joined efforts and scientific resources from both engineers and clinical medical doctors, paving the way for an accelerated transition from a traditional to a more enhanced analysis and appreciation of BAS.