PatentAuction.com lists patented inventions available for sale or licensing. Inventors can list their patented inventions (or patent pending) for sale. New inventions for sale are added on a daily basis !
Home    List your patent    My account     Help     Support us    

Enhance the Accuracy of Speech Recognition

[Category : - Audio - Video]
[Viewed 131 times]

Enhance Speech Recognition Accuracy

Introducing a groundbreaking innovation in speech recognition technology. This patent presents a method and apparatus for utilizing multiple automatic speech recognizers (ASRs) with diverse language and acoustic models to significantly improve the accuracy of speech recognition systems.

Key Features:
- Multi-ASR Integration: Employs multiple ASRs in tandem, each with unique language and acoustic models, to process speech input.
- Confidence Score Optimization: Each ASR generates a confidence score for its recognition output. The system compares these scores to select the most accurate transcription.
- Dynamic Model Selection: The system adapts to varying speech inputs by selecting the most appropriate ASR model based on the confidence scores and input characteristics.

Applications:
- Mobile Devices: Enhances voice command accuracy in smartphones and tablets, improving user experience.
- Desktop Applications: Improves speech-to-text functionalities in desktop software, aiding accessibility and productivity.
- Embedded Systems: Optimizes speech recognition in embedded devices, such as smart home assistants and automotive systems.
- Real-Time Communication: Facilitates accurate transcription in live communication platforms, ensuring clear and reliable interaction.

This patented technology offers a unique opportunity to integrate advanced speech recognition capabilities into a wide range of applications. By leveraging multiple ASRs, it ensures higher accuracy and adaptability to diverse speech inputs, setting your products apart in the competitive market.

A system configured to use multiple automatic speech recognizers (ASRs) for language and acoustics - a KEY process used to increase the accuracy of speech recognition for mobile, desktop, and other applications.







[ Home | List a patent | Manage your account | F.A.Q.|Terms of use | Contact us]

Copyright PatentAuction.com 2004-2017
Page created at 2025-07-03 6:07:02, Patent Auction Time.