Distributed Speech Recognition Using One Way Communication

• US 20110288857A1
• Filed: 08/02/2011
• Published: 11/24/2011
• Est. Priority Date: 08/29/2008
• Status: Active Grant

• Current Assignee: Solventum Intellectual Properties Company (Solventum Corp.)
• Original Assignee: Multimodal Technologies Incorporated (3M Company)
• Inventors: Carraux, Eric, Koll, Detlef
• Granted Patent: US 8,249,878 B2
• Time in Patent Office: Days
• US Class Current: 704/201
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: G10L 15/32 Multiple recognisers used in sequence or in parallel; Score combination systems therefor, e.g. voting systems

Abstract

A speech recognition client sends a speech stream and control stream in parallel to a server-side speech recognizer over a network. The network may be an unreliable, low-latency network. The server-side speech recognizer recognizes the speech stream continuously. The speech recognition client receives recognition results from the server-side recognizer in response to requests from the client. The client may remotely reconfigure the state of the server-side recognizer during recognition.

Claims

1. A computer-implemented method comprising:

(A) at a client, transmitting a speech stream and a control stream to a speech recognition server using a Hypertext Transfer Protocol (HTTP) having a first timeout period;

(B) at the speech recognition server, using an automatic speech recognition engine to initiate recognition of the speech stream;

(C) at the client, transmitting a first request for a speech recognition result to the server using HTTP; and

(D) at the server, transmitting a notification to the client indicating that no speech recognition results have become available within a second timeout period that differs from the first timeout period; and

(E) at the client, in response to receiving the notification, transmitting a second request for the speech recognition result to the server using HTTP.

2. The method of claim 1, further comprising:

(F) at the server, recognizing a first portion of the speech stream to produce a first speech recognition result; and

(G) transmitting the first speech recognition result to the client using HTTP in response to the second request.

3. The method of claim 1, wherein (G) comprises:

(G)(1) determining whether any speech recognition results are available;

(G)(2) if no speech recognition results are available, returning to (G)(1);

(G)(3) otherwise, transmitting the first speech recognition result to the client.

4. The method of claim 3, wherein the server performs (F) and (G) in parallel.

5. The method of claim 1, wherein (A) comprises transmitting the speech stream and the control stream using a Hypertext Transfer Protocol over Secure Socket Layer (HTTPS), and wherein (C) comprises transmitting the first request using HTTPS.

6. A system comprising a client device and a speech recognition server:

wherein the client device comprises;

means for transmitting a speech stream and a control stream to a speech recognition server using a Hypertext Transfer Protocol (HTTP) having a first timeout period;

means for transmitting a first request for a speech recognition result to the server using HTTP; and

wherein the speech recognition server comprises;

means for using an automatic speech recognition engine to initiate recognition of the speech stream;

means for transmitting a notification to the client indicating that no speech recognition results have become available within a second timeout period that differs from the first timeout period; and

wherein the client further comprises means, responsive to receipt of the notification, for transmitting a second request for the speech recognition result to the server using HTTP.