A Rejection Model Based on Multi-Layer Perceptrons for Mandarin Digit Recognition

High performance Mandarin digit recognition (MDR) is much more difficultto achieve than its English counterpart, especially on inexpensivehardware implementation. In this paper, a new Multi-Layer Perceptrons(MLP) based postprocessor, an a posteriori probability estimator,is presented and used for the rejection model of the hidden Markov model (HMM) Viterbi decoding inthe speaker independent Mandarin digitrecognition system based on hidden Markov model (HMM). Poor utterances, which are recognized by HMMs buthave low a posteriori probability, will be rejected. After rejectingabout 4.9% of the tested utterances, the MLP rejection model can boostthe digit recognition accuracy from 97.1% to 99.6%. The performance isbetter than those rejection models based on linear discrimination,likelihood ratio or anti-digit.