Automatic speech segmentation using neural tree networks

Segmentation of speech into sub-word acoustic units using neural tree networks (NTNs) is presented. NTN is a hierarchical classifier that combines the properties of both decision trees and feedforward neural networks. The number of sub-word acoustic units in a given speech segment may or may not be known to the segmentation algorithm. Both these varieties of speech segmentation problems are addressed. The performance of the speech segmentation algorithm using NTN is compared to that obtained using hidden Markov models (HMMs) and dynamic programming-based approach proposed elsewhere