Low-power automatic speech recognition device
AssigneeMassachusetts Institute of Technology
InventorsMichael R. Price, James R. Glass, Anantha P. Chandrakasan
Life of this patent
Estimating the term…
Abstract
A decoder comprises a feature extraction circuit for calculating one or more feature vectors; an acoustic model circuit coupled to receive one or more feature vectors from said feature extraction circuit and assign one or more likelihood values to the one or more feature vectors; a memory for storing states of transition of the decoder; and a search circuit for receiving an input from said acoustic model circuit corresponding to the one or more likelihood values based upon the one or more feature vectors, and for choosing states of transition from the memory based on the input from said acoustic model.
Claims13
A decoder comprising:
- a feature extraction circuit for calculating one or more feature vectors;
- an acoustic model circuit coupled to receive one or more feature vectors from said feature extraction circuit and assign one or more likelihood values to the one or more feature vectors;
- a memory for storing states of transition of the decoder;
- and
- a search circuit for receiving an input from said acoustic model circuit corresponding to the one or more likelihood values based upon the one or more feature vectors, and for choosing states of transition from the memory based on the input from said acoustic model circuit by performing a Viterbi search on an active state list, wherein a state lattice generated by the search is converted to an on-chip word lattice, wherein an arc in the word lattice represents a chain of one or more arcs in the state lattice.
Dependencies are inferred from the claim text.