Back to results
US 12,205,610 B2Granted
Speech imaging apparatus, speech imaging method and program
AssigneeNippon Telegraph and Telephone Corporation
InventorsYoko Ishii, Momoko Nakatani, Ai Nakane, Asuka Ono, Yumiko Matsuura
Life of this patent
Estimating the term…
Abstract
An utterance imaging device reduces a cognitive load on utterance details by extracting some character strings from character strings recognized as sounds from an utterance, acquiring an image based on the some character strings, and outputting the image to a position corresponding to a speaker related to the utterance.
Claims13
An utterance imaging device comprising:
- a processor;
- and
- a memory that includes instructions, which when executed, cause the processor to execute:
- determining positions of a plurality of speakers;
- detecting sounds from at least one speaker of the plurality of speakers;
- determining a closest speaker to the detected sounds, the closest speaker being determined from among the plurality of speakers;
- determining sounds from the closest speaker;
- extracting from the sounds from the closest speaker, using machine learning, a keyword from character strings recognized as sounds from an utterance, the character strings recognized as sounds comprising a vector space of character strings constructed using machine learning, the keyword having a relevance to a feature common to all of the plurality of speakers;
- acquiring an image based on the extracted keyword;
- and
- outputting the image to a position corresponding to the closest speaker related to the utterance, wherein the outputting the image includes outputting the image in a state where the image is rotated to align with the determined position of a speaker other than the closest speaker from among the plurality of speakers.
8 more elements
An utterance imaging method comprising, at a computer:
- determining positions of a plurality of speakers;
- detecting sounds from at least one speaker of the plurality of speakers;
- determining a closest speaker to the detected sounds, the closest speaker being determined from among the plurality of speakers;
- determining sounds from the closest speaker;
- extracting from the sounds from the closest speaker, using machine learning, a keyword from character strings recognized as sounds from an utterance, the character strings recognized as sounds comprising a vector space of character strings constructed using machine learning, the keyword having a relevance to a feature common to all of the plurality of speakers;
- acquiring an image based on the extracted keyword;
- and
- outputting the image to a position corresponding to the closest speaker related to the utterance, wherein the outputting the image includes outputting the image in a state where the image is rotated to align with the determined position of a speaker other than the closest speaker from among the plurality of speakers.
5 more elements
A system comprising:
- at least one sound collecting device configured to detect sounds;
- a display device configured to display at least one image;
- an utterance imaging device connected to each of the at least one sound collecting device and the display device, the utterance imaging device configured to:
- determine positions of a plurality of speakers;
- receive sounds from at least one speaker of the plurality of speakers, the sounds being detected by the at least one sound collecting device;
- determine a closest speaker to the detected sounds, the closest speaker being determined from among the plurality of speakers;
- determine sounds from the closest speaker;
- extract from the sounds from the closest speaker, using machine learning, a keyword from character strings recognized as sounds from an utterance, the character strings recognized as sounds comprising a vector space of character strings constructed using machine learning, the keyword having a relevance to a feature common to all of the plurality of speakers;
- acquire an image based on the extracted keyword;
- output the image to a position corresponding to the closest speaker related to the utterance, wherein outputting the image includes outputting the image in a state where the image is rotated to align with the determined position of a speaker other than the closest speaker from among the plurality of speakers;
- transmit the output image to the display device;
- and
- cause the display device to display the output image.
10 more elements
Dependencies are inferred from the claim text.