San Diego, CA (OCT 1, 2002) — LumenVox, a provider of speech recognition technologies announced a new training system, that when using their Speech Engine, dramatically enhances the accuracy of the application.
LumenVox has developed new acoustic models, based on task-dependant data to improve accuracy. This new training system will accept virtually any speech data with its transcription as input, and produce trained, effective models tailored to a company's requirements. LumenVox's Speech Engine can automatically create the speech data files from live customer calls and includes tools to simplify the transcription process.
This new training system means that LumenVox can now produce acoustic models specifically adapted for each individual client's environment, thus improving recognition accuracy and customer satisfaction. In particular, client data received by the recognition engine is sent to LumenVox, where the data is fed into the system, resulting in sharper, customized acoustic models. As the exchange of data and models continues, the client receives more accurate performance from the Speech Engine. The better matched and more data used, the better the resulting system. In time, this training system will be available to the developer and be bundled in LumenVox's suite of speech recognition software. The training system will further enhance the already existing and powerful tuning system.
For additional information or to utilize this new training system, simply call a LumenVox representative at: (877) 977-0707.
© 2017 LumenVox, LLC. All rights reserved.