Find a company or group of experts with the skills listed below:
Technology
Technology that can perform voice recognition by receiving streaming-type voice signals that are input in real time.
Mandatory
- Voice recognition processed in a local environment without using a server
- The technology to be secured is speed-oriented and must be processed immediately in a PC environment without an additional server for calculation.
- Real-time streaming input audio voice recognition
- The input voice is provided in the form of real-time streaming, and the delay between the input and the result should be minimized by continuously performing the voice recognition process before the complete sentence is made.
- Language model (LM) customization function support
- The technology needs to be applied to an environment in which a large number of unusual words and new words occur, such as streaming broadcasting. Accordingly, the language model should be able to add new words, etc., as needed, without learning.
Optional
- Mobile environment on-device voice recognition
- AOS / iOS device must be able to perform speech recognition without an additional server
- Speech recognition robustness in a noisy environment
- The voice recognition environment to which the technology is applied must function in various noise environments such as BGM and machine sounds.
Contact
- Hye-young Jeong (hyjeong@smilegate.com)
- Please send us a rough proposal and we will contact you after arranging the meeting separately.
