Word boundary detection with mel-scale frequency bank in noisy environment

This paper addresses the problem of automatic word boundary detection in th e presence of noise. We first propose an adaptive time-frequency (ATF) para meter for extracting both the time and frequency features of noisy speech s ignals. The ATF parameter extends the TF parameter proposed by Junqua et al . from Single band to multiband spectrum analysis, where the frequency band s help to make the distinction of speech and noise signals clear. The ATF p arameter can extract useful frequency information by adaptively choosing pr oper bands of the mel-scale frequency bank. The ATF parameter increased the recognition rate by about 3% of a TF-based robust algorithm which has been shown to outperform several commonly used algorithms for word boundary det ection in the presence of noise. The ATF parameter also reduced the recogni tion error rate due to endpoint detection to about 20%. Based on the ATF pa rameter, we further propose a new word boundary detection algorithm by usin g a neural fuzzy network (called SONFIN) for identifying islands of word si gnals in noisy environment. Due to the self-learning ability of SONFIN, the proposed algorithm avoids the need of empirically determining thresholds a nd ambiguous rules in normal word boundary detection algorithms. As compare d to normal neural networks, the SONFIN can always find itself an economic network size in high learning speed. Our results also showed that the SONFI N's performance is not significantly affected by the size of training set. The ATF-based SONFIN achieved higher recognition rate than the TF-based rob ust algorithm by about 5%. It also reduced the recognition error rate due t o endpoint detection to about 10%, compared to an average of approximately 30% obtained with the TF-based robust algorithm, and 50% obtained with the modified version of the Lamel ct al. algorithm.