ISM
2008
IEEE
14 years 6 months ago
2008
IEEE
We propose a multimodal speaker segmentation algorithm with two main contributions: First, we suggest a hidden Markov model architecture that performs fusion of the three modaliti...