Semi-blind speech extraction for robot using visual information and noise statistics

In this paper, speech recognition accuracy improvement is addressed for ICA-based multichannel noise reduction in spoken-dialogue robot. First, a new permutation solving method using a probability statistics model is proposed for realistic sound mixtures consisting of point-source speech and diffuse noise. Next, to achieve high recognition accuracy for the early utterance of the target speaker, we introduce a new rapid ICA initialization method combining robot video information and a prestored initial separation filter bank. From this image information, an ICA initial filter fitted to the user´s direction can be used to save the user´s first utterance. The experimental results show that the proposed approaches can markedly improve the word recognition accuracy.