An Improved VTS Feature Compensation using Mixture Models of Distortion and IVN Training for Noisy Speech Recognition | IEEE Journals & Magazine | IEEE Xplore