1. On the use of scalar quantization for fast HMM computation. ICASSP95, vol. 1, p 213-216.
2. Vector quantization for the efficient computation of continuous density likelihoods. ICASSP93, vol. 2, p 692-669.
3. Efficient speech recognition using subvector quantization and discrete-mixture HMMs. ICASSP99, SP18-4.
4. Fast output probability calculation by parameter scalar quantization and dimension-independent operation by mixed distribution HMM. Tech Rep IEICE. 1995;SP95-22.