A probabilistic framework for feature-based speech recognition

Most current speech recognizers use an observation space which is based on a temporal sequence of “frames” (e.g. Mel-cepstra). There is another class of recognizer which further processes these frames to produce a segment-based network, and represents each segment by fixed-dimensional “features”. In such feature-based recognizers, the observation space takes the form of a temporal network of feature vectors, so that a single segmentation of an utterance uses a subset of all possible feature vectors. In this paper, we examine a maximum a-posteriori decoding strategy for feature-based recognizers and develop a normalization criterion that is useful for a segment-based Viterbi or A* search. We report experimental results for the task of phonetic recognition on the TIMIT corpus, where we achieved context-independent and context-dependent (using diphones) results on the core test set of 64.1% and 69.5% respectively