Improved response, vocabulary and voice recognition accuracy
Advanced Media Corporation (Head office: Toshima-ku, Tokyo; CEO: Kazuyuki Hasegawa; hereinafter, Advanced Media) has announced the release of its iPhone app.
AmiVoice®We have completed development of two types of voice recognition engine: a DSR* version that realizes distributed voice recognition, and a standalone version that realizes voice recognition on mobile phones.
The DSR version uses the TCP/IP communication protocol, which was difficult to use on regular mobile phones due to its specifications, and performs distributed voice recognition between the iPhone and the server, improving response times.The standalone version also takes advantage of the iPhone's large application size, available memory size, and development environment that makes use of its high-performance CPU, and has succeeded in improving response times, supported vocabulary, and recognition accuracy compared to conventional mobile phones, making it possible to support solutions that were previously not possible on standalone mobile phones.
Advanced Media will develop these iPhone apps in the future.AmiVoice®Using the voice recognition engine, we plan to bring to market voice-activated email input services, voice-activated games, English conversation learning applications, etc. We also plan to sell our voice recognition library to application development companies.
Advanced Media for iPhoneAmiVoice®Starting with the launch of the DSR version of the speech recognition engine and standalone applications and services, we will continue to actively pursue support for new mobile terminals and various devices that require human-centric voice input interfaces.
Above
*DSR: Distributed Speech Recognition
This is a distributed voice recognition system in which the terminal device is equipped with a voice recognition pre-processing function and shares the processing with the server. Within the terminal device, the voice is converted into "feature data" necessary for voice recognition processing, and then sent to the server for voice recognition processing. The "feature data" sent from the terminal device is small, less than 1/30 of the voice data, and even on current mobile phone networks, there are no packet drops and it is sent to the server at a speed that does not cause stress to the user, so even though it is a small device, it can perform advanced and highly accurate voice recognition on a par with that of a PC.
[Inquiries regarding this matter]
[Press relations]
Corporate Planning Department Public Relations Team
TEL：03-5949-2007
FAX: 03-5958-1032
Email:press@advanced-media.co.jp
【customer】
Kaneda, Sales Group, Solutions Division
TEL: 03-5958-1091
FAX: 03-5958-1033
Email:info@advanced-media.co.jp
[Copyright/Trademark]
Copyright 2008 Advanced Media, Inc.
AmiVoice®and logo are registered trademarks of Advanced Media.
Company names and product names mentioned are registered trademarks and trademarks of their respective companies.
FOMA® is a registered trademark of NTT DoCoMo.
*This is an example of how to use the DSR version and the standalone version. The actual screen may differ. ↓
