Glottal Inverse Filtering and Its Application in the Automatic Classification of Diseases from Speech
摘要
Glottal inverse filtering is a technology used to estimate the source of voiced speech, the glottal volume velocity waveform, from the speech pressure signal recorded by a microphone. Based on the estimated glottal waveform, speech can be characterized using various glottal parameters. Glottal parameters can be used with machine learning and deep learning classifiers to automatically detect diseases and classify the severity level of diseases from speech signals. This chapter provides a short review of glottal inverse filtering methods, glottal parameters, and their use in the automatic classification of diseases from speech.