Solutions Showcase
Voice RecognitionVoice Recognition
For novice mobile users and elderly peoples, it is often annoying to input long email or fulfil many entry fields of input form by using mobile devices without enough size keyboard.
In order to propose a solution to this issue, Aplix has developed demonstration of voice recognition for embedded system inconjuction with Advanced Media Inc.
This demonstration was developed by combinating the JBlend™ platform and the client portion of the AmiVoiceR DSR client-server distributed speech recognition (DSR) engine from Advanced Media, Inc. The system enables Java applications running on a variety of remote devices to leverage the speech recognition engine in a client-server framework.
AmiVoice DSR is designed so that the client performs only audio analysis, with the main recognition tasks being performed on the server side. Rather than sending raw audio to the server, the client sends voice characteristics data, compressed to around 1/30th the size of the original data. Client-server communication is thereby made possible in a narrowband environment typical of mobile devices, enabling near real-time recognition.
Aplix has extended the class libraries so that this AmiVoice DSR system can be used by Java applications. Vendors will be able to create exciting new user-friendly interfaces for their mobile devices, freeing users from having to enter text on a keypad.
