News Center

What problems do you often encounter in your meeting room now?

1. Handwritten whiteboard: low grade, handwritten data cannot be saved or printed directly.

   

2. Handwritten record: slow speed, easy to miss or record errors, unable to record voice and image for the meeting.

   

3. Projection screen: the image is blurred and limited by the light; the projector is greatly affected by the environment when it is used (turn off the light) and its service life is limited. Before the meeting, special personnel shall be assigned to carry out complex meeting preparation and projector position debugging

   

4. Dull meetings reduce the efficiency of the meeting.

   

So the question is, what kind of meeting system does modern office need?

The smart speech recognition conference system of Sofitel can meet the multiple requirements of conference equipment, which is higher, more tonal, more fun, more intelligent and more efficient.

A brief introduction of smart speech recognition conference system

· The speech recognition technology of S-TRACK adopts the advanced recognition framework and uses a large number of speech data for modeling, which greatly improves the speech recognition rate and transcribing speed of the system.

· The system adopts the international leading voice language integrated modeling method, combined with Chinese Semantic Intelligent error correction, and the recognition accuracy of Chinese Mandarin is more than 95%.

Conference system architecture of smart speech recognition based on S-TRACK

System scheme features

1. Real time transcribe live broadcast: the speech is converted into text in real time and output according to the role of the speaker, which can be displayed on the screen;

2. Automatic role separation: automatically distinguish the role of the speaker, and the voice transcription recognition results automatically correspond to the corresponding role;

3. Real time correction of results: the results written in the meeting can be modified and replaced in real time to improve the accuracy of recognition results;

4. Support personalized hot words: upload some unusual words to the recognition engine to improve the recognition accuracy of professional terms;

5. Customized speech model: customized speech recognition model for customer's use region and business scope to improve the accuracy of accent and business text recognition;

6. Offline voice recognition server: safe and secure;

7. Audio: advanced audio processing chip is used, with powerful digital audio processing ability; supporting software can set its processing function through Ethernet, 48K sampling rate can effectively restore the real audio quality, and the sound effect is real and clear.

Program advantage