System And Method For Leaving And Transmitting Speech Messages

• US 20110119053A1
• Filed: 03/18/2010
• Published: 05/19/2011
• Est. Priority Date: 11/13/2009
• Status: Active Grant

• Current Assignee: Industrial Technology Research Institute
• Original Assignee: Industrial Technology Research Institute
• Inventors: Chien, Shih-Chieh, Chang, Hsin-Chang, Kuo, Chih-Chung, Chiu, Chung-Jen
• Granted Patent: US 8,660,839 B2
• Time in Patent Office: Days
• US Class Current: 704/201
• CPC Class Codes: G10L 15/19 Grammatical context, e.g. disambiguation of the recognition hypotheses based on word sequence rules
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: H04L 51/10 Multimedia information
• CPC Class Codes: H04M 2250/74 with voice recognition means voice activated dialling H04M1/271; voice recognition algorithms G10L15/00
• CPC Class Codes: H04M 3/5322 for recording text messages
• CPC Class Codes: H04M 3/533 Voice mail systems

Abstract

A system for leaving and transmitting speech messages automatically analyzes input speech of at least a reminder, fetches a plurality of tag informations, and transmits speech message to at least a message receiver, according to the transmit criterions of the reminder. A command or message parser parses the tag informations at least including at least a reminder ID, at least a transmitted command and at least a speech message. The tag informations are sent to a message composer for being synthesized into a transmitted message. A transmitting controller controls a device switch according to the reminder ID and the transmitted command, to allow the transmitted message send to the message receiver via a transmitting device.

Claims

1. A system for leaving and transmitting speech messages, comprising:

a command or message parser for parsing at least an input speech and outputting a plurality of tag information, said plurality of tag information at least including at least a reminder identity (ID), at least a transmitted command and at least a message speech;

a message composer connected to said command or message parser for composing said plurality of tag information into a transmitted message speech;

at least a message transmitting device; and

a transmission controller connected to said command or message parser, based on said at least a reminder ID and said at least a transmitted command, controlling a device switch so that said transmitted message speech being transmitted by a message transmitting device of said at least a message transmitting device to at least a message receiver.

2. The system as claimed in claim 1, wherein said command or message parser identifies from said at least an input speech said at least a reminder ID, and based on a given grammar and a speech confidence measure, identifies at least a command vocabulary segment and at least a filler segment with at least a phonetic.

3. The system as claimed in claim 2, wherein said at least a filler segment is distinguished into at least a message filler segment or at least a garbage filler segment.

4. The system as claimed in claim 3, wherein said command or message parser identifies from said at least a command vocabulary segment at least a transmitted command, based on said at least a message filler segment, said at least a message speech is extracted from said at least an input speech.

5. The system as claimed in claim 1, wherein said at least a transmitted command further includes at least a target message receiver, when to transmit said at least a speech message to said at least a target message receiver, and in what transmitting manner said at least a speech message is transmitted to said at least a target message receiver.

6. The system as claimed in claim 1, said system is a system with both one side or multiple sides transmitting and feedback capability.

7. The system as claimed in claim 1, wherein said command or message parser further includes:

a speech content extractor for receiving said at least an input speech and extracting from said input speech said reminder ID, mix type text and information of said message speech, said mix type text being mix text information of vocabulary and phonetics corresponding to said input speech; and

a text content analyzer for analyzing said mix type text to obtain said at least a transmitted command.

8. The system as claimed in claim 1, wherein said command or message parser further includes a confirmation interface, said confirmation interface executes a confirmation operation to confirm the correctness of parsed said plurality of tag information.

9. The system as claimed in claim 1, wherein said message composer selects a compose template from a language generation template database for performing sentence composition to compose said transmitted message speech.

10. The system as claimed in claim 1, wherein when a transmitting condition of said at least a transmitted command is not satisfied and is unable to accomplish transmitting, said transmission controller sets said at least a message transmitting device as a system preset transmitting manner and uses another message transmitting device of said at least a message transmitting device to transmit said transmitted message speech.

11. The system as claimed in claim 7, wherein said text content analyzer further includes:

a concept sequence restructure for re-editing said mix type text to generate a plurality of concept sequences; and

a concept sequence selection for computing a total score corresponding to each of said plurality of concept sequences and selecting at least an optimal concept sequence formed by a semantic frame;

wherein said total score corresponding to each concept sequence is computed based on a confidence and a concept score corresponding to said concept sequence.

12. The system as claimed in claim 1, wherein said message composer further includes:

a language generator for receiving said at least a reminder ID, said at least a transmitted command and said at least a message speech, and selecting a compose template to generate a speech signal; and

a speech synthesis for composing said speech signal and said at least a message speech into said transmitted message speech.

13. The system as claimed in claim 12, wherein said compose template is selected from a language generation template database, said language generation template database includes a plurality of transmitted message compose templates or a plurality of feedback message compose templates, or a plurality of transmitted message compose templates and a plurality of feedback message compose templates.

14. A method for leaving and transmitting speech messages, comprising:

parsing at least an input speech of at least a reminder into a plurality of tag information, said plurality of tag information at least including at least a reminder identity (ID), at least a transmitted command and at least a message speech;

composing said plurality of tag information into a transmitted message speech; and

based on said at least a reminder ID and said at least a transmitted command, controlling a device switch so that said transmitted message speech being transmitted through a message transmitting device of said at least a message transmitting device to said at least a message receiver.

15. The method as claimed in claim 14, further including:

identifying said at least a reminder ID from said at least an input speech;

based on a given grammar and a speech confidence measure, identifying at least a text command segment and at least a filler segment; and

from said at least a text command segment, obtaining said at least a transmitted command, and based on said at least a filler segment, obtaining said at least a message speech.

16. The method as claimed in claim 14, further including:

based on said at least a reminder ID, said at least a transmitted command and said at least a message speech, selecting a compose template from a language generation template database to generate a speech signal; and

composing said speech signal and said at least a message speech into said transmitted message speech.

17. The method as claimed in claim 14, wherein said at least a transmitted command further includes at least a transmitting condition, when said at least a transmitting condition is met, a specified reminder manner is used to accomplish transmitting said transmitted message speech.

18. The method as claimed in claim 17, wherein when a transmitting condition of said at least a transmitting condition is not met and transmitted unsuccessfully, a system preset transmitting manner is used to transmit said transmitted message speech.

19. The method as claimed in claim 14, wherein parsing said at least an input speech of said at least a reminder further includes:

extracting said reminder ID, mix type text and information from said at least an input speech, said mix type text being mix text information of vocabulary and phonetics corresponding to said at least an input speech; and

analyzing said mix type text to obtain said at least a transmitted command.

20. The method as claimed in claim 19, wherein analyzing said mix type text further includes:

through a concept composer grammar, an example concept sequence language material databank and a message or garbage grammar to re-edit said mix type text to generate a plurality of concept sequences, then computing a corresponding confidence for each of said plurality of concept sequences; and

computing a corresponding concept score of each concept sequence, and based on said corresponding confidence and said corresponding concept score of each concept sequence, computing a total score corresponding to each concept sequence, and selecting an optimal concept sequence formed by a semantic frame.

21. The method as claimed in claim 14, further including:

before transmitting said transmitted message speech, using a confirmation interface to execute at least a confirmation operation to confirm correctness of said plurality of tag information or said transmitted message speech.

22. The method as claimed in claim 20, wherein a total score corresponding to each of said plurality concept sequences is a weighted sum of said corresponding concept score and said corresponding confidence.

23. The method as claimed in claim 17, further including:

determining whether said at least a transmitting condition of said at least a transmitted command is being met via at least a sensor device.

24. The method as claimed in claim 18, wherein said transmitted message speech is a feedback message.