Comparing hysteresis comparator and RMS threshold methods for automatic single cough segmentations

Research on diagnosing diseases based on voice signals is rapidly increasing, including cough-related diseases. When training the cough sound signals into deep learning models, it is necessary to have a standard input by segmenting several cough signals into individual cough signals. Segmenting coughs could also be used to monitor trends of cough-related disease by counting the number of coughs. Previous research has been developed to segment cough signals from non-cough signals. This research evaluates the segmentation methods of several cough signals from a single audio file into several audio files containing a single file. We evaluate three different methods, including manual segmentation as a baseline and two automatic segmentation methods: hysteresis comparator and root mean square (RMS) methods. The results by two automatic segmentation methods obtained precisions of 73% (hysteresis) and 70% (RMS) compared to 49% by manual segmentation. The agreements of listening tests to count the number of correct single-cough segmentations show fair and moderate correlations for automatic segmentation methods and are comparable with manual segmentation.