Extension of Pre-Image Speech De-Noising by Voice Activity Detection Using a Bone Conductive Microphone

In this paper, we use voice activity detection to improve the de-noising ability of the previously proposed pre-image iteration speech enhancement method. We use a speech database consisting of two-channel recordings where the audio signal is recorded by both a bone conductive microphone and a close-talking microphone. The bone channel is used for voice activity detection as it can be assumed to be robust against environmental noise. The pre-image iteration method is prone to residual noise around speech components we use the voice activity detection to remove this noise. The approach is evaluated using objective quality measures of the PEASS toolbox and shows an increase of the de-noising capability compared to the original method.