18268536. SPEAKER IDENTIFICATION APPARATUS, METHOD, AND PROGRAM simplified abstract (NEC Corporation)

From WikiPatents
Jump to navigation Jump to search

SPEAKER IDENTIFICATION APPARATUS, METHOD, AND PROGRAM

Organization Name

NEC Corporation

Inventor(s)

Qiongqiong Wang of Tokyo (JP)

Takafumi Koshinaka of Tokyo (JP)

SPEAKER IDENTIFICATION APPARATUS, METHOD, AND PROGRAM - A simplified explanation of the abstract

This abstract first appeared for US patent application 18268536 titled 'SPEAKER IDENTIFICATION APPARATUS, METHOD, AND PROGRAM

Simplified Explanation

The abstract describes a system that selects a subset of speakers based on certain attributes and uses this subset to identify the speaker of a test utterance. The identification is done by extracting features from the test utterance and comparing them to the subset of speakers.

  • The system selects speakers based on specific attributes from a larger set of speaker information.
  • It determines a subset of a speech model using the selected speakers.
  • The system identifies the speaker of a test utterance by comparing its features to the subset of speakers in the speech model.

Potential applications of this technology:

  • Speaker identification in security systems: This technology can be used to accurately identify speakers in security systems, such as access control or surveillance systems.
  • Voice-based authentication: It can be used for voice-based authentication in various applications, such as banking or personal device access.
  • Forensic analysis: This technology can assist in forensic analysis by identifying speakers from recorded conversations or phone calls.

Problems solved by this technology:

  • Accurate speaker identification: The system improves the accuracy of speaker identification by selecting a subset of speakers based on specific attributes, reducing the potential pool of speakers for comparison.
  • Efficient processing: By using a subset of speakers, the system can process the test utterance more efficiently, reducing computational resources required.

Benefits of this technology:

  • Enhanced security: By accurately identifying speakers, this technology can enhance security measures in various applications, preventing unauthorized access or fraudulent activities.
  • Improved user experience: Voice-based authentication can provide a convenient and seamless user experience, eliminating the need for passwords or PINs.
  • Time-saving in forensic analysis: This technology can save time in forensic analysis by quickly identifying speakers from recorded conversations, aiding in investigations.


Original Abstract Submitted

A speaker subset selection means selects speakers corresponding to an attribute from subset information of an entire speaker to determine a subset of a speech model from which test utterance is identified. A speaker identification means identifies a speaker of the test utterance from a subset of the determined speech model based on features extracted from the test utterance.

Retrieved from "http://wikipatents.org/index.php?title=18268536._SPEAKER_IDENTIFICATION_APPARATUS,_METHOD,_AND_PROGRAM_simplified_abstract_(NEC_Corporation)&oldid=31892"