Back to results
US 6,789,063 B1Granted
Acoustic modeling using a two-level decision tree in a speech recognition system
AssigneeIntel Corporation
InventorYonghong Yan
Life of this patent
Estimating the term…
Abstract
In some embodiments, the invention involves receiving phonetic samples and assembling a two-level phonetic decision tree structure using the phonetic samples. The decision tree has multiple leaf node levels each having at least one state, wherein a least one node in a second level is assigned a Gaussian of a node in the first level, but the at least one node in the second level has a weight computed for it.
Claims21
A method comprising:
- receiving acoustic samples;
- and
- assembling a two-level decision tree structure using the acoustic samples, wherein the two-level decision tree structure comprises at least two leaf node levels each having at least one state, and all the states in the same leaf node share a Gaussian pool.
A method comprising:
- receiving acoustic samples;
- and
- assembling a two-level decision tree structure using the acoustic samples, wherein the two-level decision tree structure comprises at least two leaf node levels each having at least one state, wherein at least one node in a second leaf node level is assigned a Gaussian component of a node in a first leaf node level, but the at least one node in the second level has at least one weight computed for it.
An article comprising:
- a storage medium to hold instructions which when executed cause a computer to performing the following method:
- receiving acoustic samples;
- and
- assembling a two-level decision tree structure using the acoustic samples, wherein the two-level decision tree structure comprises at least two leaf node levels each having at least one state, and all the states in the same leaf node share a Gaussian pool.
1 more element
An article comprising:
- a storage medium to hold instructions which when executed cause a computer to performing the following method:
- receiving acoustic samples;
- and
- assembling a two-level decision tree structure using the acoustic samples, wherein the two-level decision tree structure comprises at least two leaf node levels each having at least one state, wherein at least one node in a second leaf node level is assigned a Gaussian component of a node in a first leaf node level, but the at least one node in the second level has at least one weight computed for it.
1 more element
A two-level decision tree structure, comprising:
- a plurality of first-level leaf nodes, wherein each first-level leaf node associates with a Gaussian pool;
- and
- a plurality of second-level leaf nodes, wherein all second-level leaf nodes that derive from a first-level leaf node share a same Gaussian pool with the first-level leaf node, but each of these second-level leaf nodes has its own set of weights for each Gaussian component in the same Gaussian pool.
Dependencies are inferred from the claim text.