Fig. 9From: A Bayesian view on acoustic model-based techniques for robust speech recognitionBayesian network representation of a conditional and b combined-order HMMS (Subsection 6.9). Subfigure a is based on [26]Back to article page