18664348. BACKGROUND AUDIO IDENTIFICATION FOR SPEECH DISAMBIGUATION simplified abstract (GOOGLE LLC)
Contents
BACKGROUND AUDIO IDENTIFICATION FOR SPEECH DISAMBIGUATION
Organization Name
Inventor(s)
Jason Sanders of New York NY (US)
Gabriel Taubman of Brooklyn NY (US)
John J. Lee of Long Island City NY (US)
BACKGROUND AUDIO IDENTIFICATION FOR SPEECH DISAMBIGUATION - A simplified explanation of the abstract
This abstract first appeared for US patent application 18664348 titled 'BACKGROUND AUDIO IDENTIFICATION FOR SPEECH DISAMBIGUATION
The patent application describes a method for providing context-dependent search results based on audio streams containing user speech data and background audio.
- The method involves separating the audio stream into user speech data and background audio.
- Concepts related to the background audio are identified, and a set of terms related to these concepts is generated.
- The speech recognizer is influenced based on the terms related to the background audio to improve recognition of user speech data.
- The recognized version of the user speech data is obtained using the speech recognizer.
Potential Applications: - This technology could be used in voice-controlled devices to improve the accuracy of speech recognition in noisy environments. - It could enhance the user experience in virtual assistants by better understanding user commands in various audio settings.
Problems Solved: - Addressing the challenge of accurately recognizing user speech data in the presence of background audio. - Improving the performance of speech recognition systems in real-world scenarios.
Benefits: - Enhanced accuracy and efficiency in speech recognition tasks. - Improved user satisfaction with voice-controlled devices and applications.
Commercial Applications: "Context-Dependent Speech Recognition Technology for Enhanced User Experience in Voice-Controlled Devices"
Questions about Context-Dependent Speech Recognition Technology: 1. How does this technology differentiate between user speech data and background audio? 2. What are the potential limitations of using background audio to influence speech recognition accuracy?
Frequently Updated Research: Stay updated on advancements in speech recognition technology and the integration of context-dependent features for improved performance.
Original Abstract Submitted
Implementations relate to techniques for providing context-dependent search results. A computer-implemented method includes receiving an audio stream at a computing device during a time interval, the audio stream comprising user speech data and background audio, separating the audio stream into a first substream that includes the user speech data and a second substream that includes the background audio, identifying concepts related to the background audio, generating a set of terms related to the identified concepts, influencing a speech recognizer based on at least one of the terms related to the background audio, and obtaining a recognized version of the user speech data using the speech recognizer.
