Back to results
US 11,538,488 B2Granted
Method and system for processing speech signal
AssigneeAlibaba Group Holding Limited
InventorsShiliang Zhang, Ming Lei, Wei Li, Haitao Yao
Life of this patent
Estimating the term…
Abstract
Embodiments of the present disclosure provide methods and systems for processing a speech signal. The method can include: processing the speech signal to generate a plurality of speech frames; generating a first number of acoustic features based on the plurality of speech frames using a frame shift at a given frequency; and generating a second number of posteriori probability vectors based on the first number of acoustic features using an acoustic model, wherein each of the posteriori probability vectors comprises probabilities of the acoustic features corresponding to a plurality of modeling units, respectively.
Claims14
A method for processing a speech signal, comprising:
- processing the speech signal to generate a plurality of speech frames;
- generating a first number of acoustic features based on the plurality of speech frames using a frame shift at a given frequency;
- and
- generating a second number of posteriori probability vectors based on the first number of acoustic features, each of the posteriori probability vectors comprising probabilities of the acoustic features corresponding to a plurality of modeling units, respectively, wherein:
- generation uses an acoustic model trained by:
- acquiring a plurality of stored speech signals;
- generating a plurality of acoustic features and modeling units corresponding to the plurality of acoustic features based on the plurality of stored speech signals;
- generating, based on the plurality of acoustic features, a plurality of acoustic feature sequences, each having adjacent acoustic features;
- determining average probability vectors of the plurality of acoustic feature sequences corresponding to the plurality of modeling units, respectively;
- and
- generating the acoustic model using the acoustic feature sequences and the average probability vectors corresponding to the acoustic feature sequences;
- and
- generation further comprises:
- classifying, by the acoustic model, the first number of acoustic features into the second number of acoustic feature sequences, each having a plurality of adjacent acoustic features;
- and
- generating, by the acoustic model, the second number of posteriori probability vectors based on the second number of acoustic feature sequences, respectively.
13 more elements
An apparatus for processing a speech signal, comprising:
- a memory storing a set of instructions;
- and
- at least one processor configured to execute the set of instructions to cause the apparatus to perform:
- processing the speech signal to generate a plurality of speech frames;
- generating a first number of acoustic features based on the plurality of speech frames using a frame shift at a given frequency;
- and
- generating a second number of posteriori probability vectors based on the first number of acoustic features, each of the posteriori probability vectors comprising probabilities of the acoustic features corresponding to a plurality of modeling units, respectively, wherein:
- generation uses an acoustic model trained by:
- acquiring a plurality of stored speech signals;
- generating a plurality of acoustic features and modeling units corresponding to the plurality of acoustic features based on the plurality of stored speech signals;
- generating, based on the plurality of acoustic features, a plurality of acoustic feature sequences, each having adjacent acoustic features;
- determining average probability vectors of the plurality of acoustic feature sequences corresponding to the plurality of modeling units, respectively;
- and
- generating the acoustic model using the acoustic feature sequences and the average probability vectors corresponding to the acoustic feature sequences;
- and
- generation further comprises:
- classifying, by the acoustic model, the first number of acoustic features into the second number of acoustic feature sequences, each having a plurality of adjacent acoustic features;
- and
- generating, by the acoustic model, the second number of posteriori probability vectors based on the second number of acoustic feature sequences, respectively.
16 more elements
A non-transitory computer readable medium that stores a set of instructions that is executable by at least one processor of a computer system to cause the computer system to perform a method for processing a speech signal, the method comprising:
- processing the speech signal to generate a plurality of speech frames;
- generating a first number of acoustic features based on the plurality of speech frames using a frame shift at a given frequency;
- and
- generating a second number of posteriori probability vectors based on the first number of acoustic features, each of the posteriori probability vectors comprising probabilities of the acoustic features corresponding to a plurality of modeling units, respectively, wherein:
- generation uses an acoustic model trained by:
- acquiring a plurality of stored speech signals;
- generating a plurality of acoustic features and modeling units corresponding to the plurality of acoustic features based on the plurality of stored speech signals;
- generating, based on the plurality of acoustic features, a plurality of acoustic feature sequences, each having adjacent acoustic features;
- determining average probability vectors of the plurality of acoustic feature sequences corresponding to the plurality of modeling units, respectively;
- and
- generating the acoustic model using the acoustic feature sequences and the average probability vectors corresponding to the acoustic feature sequences;
- and
- generation further comprises:
- classifying, by the acoustic model, the first number of acoustic features into the second number of acoustic feature sequences, each having a plurality of adjacent acoustic features;
- and
- generating, by the acoustic model, the second number of posteriori probability vectors based on the second number of acoustic feature sequences, respectively.
13 more elements
Dependencies are inferred from the claim text.