Reverberation and noise robust feature enhancement using multiple inputs
Shin Jae Kang, Tae Gyoon Kang, Kang Hyun Lee, Kiho Cho, Nam Soo Kim · 2014
We propose a novel approach to feature enhancement in multi-channel scenario. Our approach is based on the interacting multiple model (IMM), which was originally developed in single-channel scenario. We extend the single-channel IMM algorithm such that it can handle the multichannel inputs under the Bayesian framework. The multichannel IMM algorithm is capable of tracking time-varying room impulse responses and background noises by updating the relevant parameters in an on-line manner. In various environmental conditions, the performance gain of the proposed method has been confirmed.