Voice recognition system having a check scheme for registration of reference data

• US 4,769,844 A
• Filed: 04/03/1987
• Issued: 09/06/1988
• Est. Priority Date: 04/03/1986
• Status: Expired due to Fees

• Current Assignee: Ricoh Company Limited
• Original Assignee: Ricoh Company Limited
• Inventors: Muroi, Tetsuya, Fujimoto, Junichiroh
• Primary Examiner(s): Miska, Vit W.
• Application Number:
  US07/033,943
• Time in Patent Office: 522 Days
• Field of Search: 381/41-43, 364/513.5
• US Class Current: 704/248
• CPC Class Codes: G10L 15/00 Speech recognition G10L17/00 takes precedence

Abstract

A voice recognition system includes a microphone for converting a voice in the form of pressure waves into an electrical voice signal and a keyboard for inputting the phonetic spelling of the voice. A check mechanism is connected to the keyboard to determine whether or not the voice has a predetermined sound element, which is difficult to be detected, such as /u/, /i/, /n/ and explosives like /p/, /t/, /k/ and /b/, at or near the end thereof. If the check mechanism has determined the presence of such a predetermined sound element, then it checks the corresponding voice pattern to see whether the corresponding component is present in the pattern or not. If not, a warning may be issued to request reinputting of the voice, thereby insuring the fact that the voice pattern to be registered as a reference voice pattern is always accurate.

Claims

1. A voice recognition system comprising:

first input means for inputting a voice in the form of pressure waves and converting said voice into an electrical signal;

processing means for processing said electrical signal to produce a voice pattern of a predetermined format;

second input means for inputting an identification information of said voice;

storing means for storing said voice pattern as a reference pattern selectively; and

checking means for checking said identification information to determine whether or not said voice has a predetermined sound element at or near the end thereof, thereby causing said voice pattern to be stored in said storing means if said voice has been found to have no predetermined sound element and then to check said voice pattern to determine whether or not said voice pattern has a corresponding component of said predetermined sound element if said voice has been found to have said predetermined sound element.

2. The system of claim 1 wherein said checking means issues a request of reinputting said voice into said first input means if said voice pattern has been found to have no corresponding component of said predetermined sound element.

3. The system of claim 1 further comprising averaging means for averaging a plurality of said voice patterns to define an average voice pattern before being stored into said storing means.

4. The system of claim 1 wherein said predetermined sound element is a selected vowel.

5. The system of claim 4 wherein said selected vowel is /u/ or /i/.

6. The system of claim 1 wherein said predetermined sound element is an explosive.

7. The system of claim 1 wherein said first input means includes an audio-electro transducer for converting a voice in the form of pressure waves into an electrical signal.

8. The system of claim 7 wherein said transducer is a microphone.

9. The system of claim 1 wherein said second input means includes a keyboard and said identification information is a phonetic spelling of said voice, whereby said keyboard converts said phonetic spelling into a string of codes.

10. The system of claim 9 wherein said checking means receives said string of codes to determine the presence or absence of said predetermined sound element in said voice.

11. The system of claim 1 further comprising:

selecting means for selecting a registration mode or a recognition mode; and

matching means for matching said voice pattern with each of said reference patterns stored in said storing means while keeping said checking means inoperative during said recognition mode, whereby one of said reference patterns which is most similar to said voice pattern is selected as an output.

12. A voice recognition system comprising:

first input means for inputting a voice in the form of pressure waves and converting said voice into an electrical signal;

processing means for processing said electrical signal to produce a voice pattern of a predetermined format;

second input means for inputting an identification information of said voice;

first storing means for storing said voice pattern as a reference pattern selectively;

second storing means for storing at least one explosive voice pattern;

checking means for checking said identification information to determine whether or not said voice has a voiceless interval near the end thereof and if so to determine whether or not said voice pattern has said voiceless interval near the end thereof; and

coupling means for coupling said voice pattern with a corresponding explosive voice pattern stored in said second storing means to define a complete voice pattern before being stored in said first storing means as a reference pattern.

13. The system of claim 12 wherein said corresponding explosive voice pattern is combined with said voice pattern with a predetermined gap therebetween.

14. The system of claim 12 wherein said first input means includes an audio-electro transducer.

15. The system of claim 14 wherein said audio-electro transducer is a microphone.

16. The system of claim 12 wherein said second input means includes a keyboard and said identification information is a phonetic spelling of said voice.

17. The system of claim 12 further comprising:

selecting means for selecting one of a first and second registration modes and a recognition mode, said selecting means causing said processing means and said second storing means to be connected to said coupling means when said first registration mode is established and causing said processing means to be connected to said second storing means when said second registration mode is established;

matching means for matching said voice pattern with each of said reference patterns stored in said first storing means while keeping said checking means inoperative during said recognition mode; and

means for determining which of said reference patterns in said first storing means is most similar to said voice pattern.