Method and system for muting classified information from an audio

Patent metadata

Field	Value
Publication number	US-10930286-B2
Application number	US-201916254387-A
Country	US
Kind code	B2
Filing date	Jan 22, 2019
Priority date	Jul 16, 2018
Publication date	Feb 23, 2021
Grant date	Feb 23, 2021

Abstract

This disclosure relates generally to a method and system for muting of classified information from an audio using a fuzzy approach. The method comprises converting the received audio signal into text using a speech recognition engine to identify a plurality of classified words from the text to obtain a first set of parameters. Further, a plurality of subwords associated with each classified word are identified to obtain a second set of parameters associated with each subword of corresponding classified word. A relative score is computed for each subword associated with the classified word based on a plurality of similar pairs for the corresponding classified word. A fuzzy muting function is generated using the first set of parameters, the second set of parameters and the relative score associated with each subword. The plurality of subwords associated with each classified word is muted in accordance with the generated fuzzy muting function.

First claim

What is claimed is:

1. A processor-implemented method for muting classified information in an audio signal, the method comprises:
receiving, an audio signal from the data source;
converting, the audio signal into text using a speech recognition engine;
identifying, a plurality of classified words from the text using a repository of predefined classified words;
obtaining, a first set of parameters associated with each classified word among the plurality of classified words;
identifying, a plurality of subwords associated with each classified word;
obtaining, a second set of parameters associated with each subword among the plurality of subwords of corresponding classified word among the plurality of classified words;
computing, a relative score for each subword associated with the corresponding classified word based on a plurality of similar pairs for the corresponding classified word, wherein computing the relative score for each subword of the corresponding classified word comprises:
obtaining, a list of words phonetically similar to the corresponding classified word using a language lexicon;
obtaining, the plurality of similar pairs by pairing each subword in the obtained list of words with the classified words at subword level;
determining, a count of the plurality of word level similar pairs;
clustering, the plurality of similar pairs into a plurality of clusters based on a predefined cluster criteria for each subword associated with the corresponding classified word, wherein each cluster among the plurality of clusters comprises a set of subword level similar pairs; and
determining, the relative score for each subword associated with the corresponding classified word, wherein the relative score is a ratio of plurality of clustered subword level similar pairs based on a predefined cluster criteria to the count of the plurality of word level similar pairs;
generating, a fuzzy muting function using the first set of parameters, the second set of parameters and the relative score associated with each subword; and
muting, the plurality of subwords associated with each classified word in accordance with the generated fuzzy muting function.

2. The method as claimed in claim 1 , wherein the first set of parameters comprises a time stamp of each classified word, a duration for each classified word occurred and a confidence score of each classified word.

3. The method as claimed in claim 1 , wherein the second set of parameters comprises a time stamp of each subword, a duration for which each subword occurred and a subword confidence score associated with each classified word.

4. An information muting system ( 102 ) for muting classified information in an audio signal, wherein the information muting system ( 102 ) comprises:
a processor ( 202 );
an Input/output (I/O) interface ( 204 ); and
a memory ( 208 ) coupled to the processor ( 202 ), the memory ( 208 ) comprising:
a conversion module ( 212 ) is configured to convert the audio signal into text using a speech recognition engine;
a classified word detection module ( 214 ) is configured to:
identify a plurality of classified words from the text using a repository of predefined classified words; and
obtain a first set of parameters associated with each classified word among the plurality of classified words;

a subwords module ( 216 ) is configured to:
identify a plurality of subwords associated with each classified word;
obtain a second set of parameters associated with each subword among the plurality of subwords of corresponding classified word among the plurality of classified words; and

compute a relative score for each identified subword associated with the classified word based on a plurality of similar pairs for the classified word, wherein the subwords module ( 216 ) is configured to compute the relative score for each subword of the corresponding classified word, by:
obtaining, a list of words phonetically similar to the corresponding classified word using a language lexicon;
obtaining, the plurality of similar pairs by pairing each subword in the obtained list of words to the classified words at subword level;
determining, a count of the plurality of word level similar pairs;
clustering, the plurality of similar pairs into a plurality of clusters based on a predefined cluster criteria for each subword associated with the corresponding classified word, wherein each cluster among the plurality of clusters comprises a set of subword level similar pairs; and
determining, the relative score for each subword associated with the corresponding classified word, wherein the relative score is a ratio of plurality of clustered subword level similar pairs based on a predefined cluster criteria to the count of the plurality of word level similar pairs;
a fuzzy muting module ( 218 ) is configured to:
generate a fuzzy muting function using the first set of parameters, the second set of parameters and the relative score associated with each subword; and
mute the plurality of subwords associated with each classified word in accordance with the generated fuzzy muting function.

5. The information muting system ( 102 ) as claimed in claim 4 , wherein the first set of parameters comprises a time stamp of each classified word, a duration of each classified word occurred and a confidence score of each classified word.

6. The information muting system ( 102 ) as claimed in claim 4 , wherein the second set of parameters comprises a time stamp of each subword, a duration of each subword occurred and a subword confidence score associated with each classified word.

7. One or more non-transitory machine readable information storage mediums comprising one or more instructions which when executed by one or more hardware processors perform actions comprising:
receiving an audio signal from the data source;
converting the audio signal into text using a speech recognition engine;
identifying a plurality of classified words from the text using a repository of predefined classified words;
obtaining a first set of parameters associated with each classified word among the plurality of classified words;
identifying a plurality of subwords associated with each classified word;
obtaining a second set of parameters associated with each subword among the plurality of subwords of corresponding classified word among the plurality of classified words;
computing a relative score for each subword associated with the corresponding classified word based on a plurality of similar pairs for the corresponding classified word, wherein computing the relative score for each subword of the corresponding classified word comprises:
obtaining a list of words phonetically similar to the corresponding classified word using a language lexicon;
obtaining the plurality of similar pairs by pairing each subword in the obtained list of words with the classified words at subword level;
determining a count of the plurality of word level similar pairs;
clustering the plurality of similar pairs into a plurality of clusters based on a predefined cluster criteria for each subword associated with the corresponding classified word, wherein each cluster among the plurality of clusters comprises a set of subword level similar pairs; and
determining the relative score for each subword associated with the corresponding classified word, wherein the relative score is a ratio of plurality of clustered subword level similar pairs based on a predefined cluster criteria to the count of the plurality of word level similar pairs;
generating a fuzzy muting function using the first set of parameters, the second set of parameters and the relative score associated with each subword; and
muting the plurality of subwords associated with each classified w

Assignees

• Tata Consultancy Services Ltd

Inventors

• Sheikh Imran Ahamad
• Kopparapu Sunil Kumar
• Vachhani Bhavikkumar Bhagvanbhai
• Garlapati Bala Mallikarjunarao
• Chalamala Srinivasa Rao

Classifications

• G10L15/222 Primary

  Barge in, i.e. overridable guidance for interrupting prompts · CPC title

• G10L15/063

  Training · CPC title

• G10L15/08 Primary

  Speech classification or search · CPC title

• G06N7/026

  Development tools for entering the parameters of a fuzzy system · CPC title

• G10L25/48

  specially adapted for particular use · CPC title