A 54-mw 3×-real-time 60-kword continuous speech recognition processor VLSI

Abstract

This paper describes a low-power VLSI chip for speaker-independent 60-kWord continuous speech recognition. We implement parallel and pipelined architecture for GMM computation and Viterbi processing. It includes a 8-path Viterbi transition architecture to maximize the processing speed and adopts tri-gram language model to improve the recognition accuracy. A two-level cache architecture is implemented for the demo system. Measured results show that our implementation achieves 25% required frequency reduction (62.5MHz) and 26% power consumption reduction (54.8mW) for 60k-Word real-time continuous speech recognition compared to the previous work. This chip can maximally process 3.02× and 2.25× times faster than real-time at 200MHz using the bigram and trigram language models, respectively.