Method and device for activating a voice-controlled function in a multi-station network through using both speaker-dependent and speaker-independent speech recognition

• US 6,792,083 B2
• Filed: 10/07/1998
• Issued: 09/14/2004
• Est. Priority Date: 10/07/1997
• Status: Expired due to Fees

• Current Assignee: US Philips Corporation (Koninklijke Philips N.V.)
• Original Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Inventors: Dams, Franciscus J. L., Hesdahl, Piet B., Van Velden, Jeroen G.
• Primary Examiner(s): Tsang, Fan
• Assistant Examiner(s): Gauthier, Gerald
• Application Number:
  US09/167,818
• Publication Number:
  US 20030147510A1
• Time in Patent Office: 2,169 Days
• Field of Search: 379/88.01, 379/142.01, 379/221.09, 379/88.02, 379/88.03, 379/88.04, 379/88.13, 704/241, 704/240, 704/253, 704/260, 704/231, 704/256
• US Class Current: 379/88.01
• CPC Class Codes: G10L 15/06 Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice G10L15/14 takes precedence
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/065 Adaptation
• CPC Class Codes: G10L 15/10 using distance or distortion measures between unknown speech and reference templates
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 2015/0631 Creating reference templates; Clustering
• CPC Class Codes: G10L 2015/228 of application context

Abstract

A voice-controlled multi-station network has both speaker-dependent and speaker-independent speech recognition. Conditionally to recognizing items of an applicable vocabulary, the network executes a particular function. The method receives a call from a particular origin and executes speaker-independent speech recognition on the call. In an improvement procedure, in case of successful determination of what has been said, a template associated to the recognized speech item is stored and assigned to the origin. Next, speaker-dependent recognition is applied if feasible, for speech received from the same origin, using one or more templates associated to that station. Further, a fallback procedure to speaker-independent recognition is maintained for any particular station in order to cater for failure of the speaker-dependent recognition, while allowing reverting to the improvement procedure.

Claims

1. A method for activating a voice-controlled function in a multi-station network by using both speaker-dependent and speaker-independent speech recognition facilities, and conditionally to recognizing one or more items or an applicable vocabulary, driving one or more network parts to activate said function, wherein said method comprises the following steps:

receiving a station-initiated call containing one or more initial speech items from the vocabulary,

executing speaker-independent recognition on said initial speech items through one or more general templates, whilst in an speech recognition improvement procedure, in case of successful ascertainment of what had been actually spoken, storing a particular speaker-specific template derived from the initial speech item so recognized and assigned to an origin of the call in question, said speaker-specific template being cyclically retained for subsequent speaker-dependent recognition of additional speech items having the same origin;

following said speech recognition improvement procedure, applying speaker-dependent recognition as an initial type of speech recognition if feasible for additional speech items received from the same origin, through one or more particular templates associated to that origin and only subsequently applying speaker-independent recognition as a fallback procedure if the recognition of the additional speech items cannot be ascertained by speaker-dependent recognition,

wherein speaker-independent recognition is a first response for new or unidentified users of the voice-controlled function, and speaker-dependent recognition based on said speech recognition improvement procedure is a first response for repeat users of the voice-controlled function, with a reversion to speaker-independent recognition if the additional speech items are not recognized.

2. The method as claimed in claim 1, wherein said origin is defined by a Calling Line Identity (CLI).

3. The method as claimed in claim 1, and providing for externally defining a speech item for which both speaker-dependent and speaker-independent recognition had been unsuccessful and/or erroneous, thereby effecting said ascertaining.

4. The method as claimed in claim 3, and allowing for then storing a particular template derived from the non-recognized speech item.

5. The method as claimed in claim 3, and allowing for then storing a general template derived from the non-recognized speech item.

6. The method as claimed in claim 1, wherein said function includes a directory search based on an identifier received in the form of speech.

7. The method as claimed in claim 1, wherein the vocabulary is predefined and finite.

8. The method as claimed in claim 1, and cyclically refreshing a set of templates originating from the same origin and representing the same speech item.

9. The method as claimed in claim 1, and treating an unidentified origin as a default origin additional to all registered origins.

10. A device being arranged for executing the method as claimed in claim 1.