Speech modeling and voiced/unvoiced/mixed/silence speech segmentation with fractionally Gaussian noise based models | IEEE Conference Publication | IEEE Xplore