| 摘要 |
PROBLEM TO BE SOLVED: To generate a speech recognition result with a high-speed response time without applying too much burden on limited computing resources of a client device.SOLUTION: A speech recognition client transmits a voice stream and a control stream in parallel to a speech recognition device on a server side via a network. The network may be a low-delay network with low reliability. The speech recognition device on the server side continuously recognizes the voice stream. The speech recognition client receives a recognition result from the recognition device on the server side responding to a request from the client. In the process of recognition, the client may reconfigure the state of the recognition device on the server side by remote. |
