Customized personal terminal device

• US 5,163,111 A
• Filed: 08/14/1990
• Issued: 11/10/1992
• Est. Priority Date: 08/18/1989
• Status: Expired due to Fees

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Akimoto, Hajime, Nakagawa, Tetsuya, Tonomura, Motonobu, Noguchi, Kouki, Baji, Toru, Masuhara, Toshiaki
• Primary Examiner(s): MacDonald, Allen R.
• Application Number:
  US07/567,010
• Time in Patent Office: 819 Days
• Field of Search: 364/513, 395/22, 395/2, 382/14-15
• US Class Current: 706/20
• CPC Class Codes: G06F 3/16 Sound input; Sound output speech processing G10L
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/187 Frequency domain transformation; Autocorrelation
• CPC Class Codes: G07C 9/25 using biometric data, e.g. fingerprints, iris scans or voice recognition
• CPC Class Codes: G10L 15/16 using artificial neural networks

Abstract

There is provided a customized personal terminal device capable of operating in response to input data peculiar to the operator, comprising a speech recognition unit for recognizing inputted speech, an image recognition unit for recognizing inputted image, and an instruction recognition unit for recognizing an inputted instruction. Neural networks are provided in at least two of the speech, image and instruction recognition units, a bus operatively connected to the respective recognition units, a processor operatively connected to the bus to perform processing upon the speech, and image and instruction recognized by the recognition units. Also, memory is operatively connected to the bus, and a control unit exercises control over information exchange between respective recognition units and the memory under the control of the processor.

Claims

1. A customised personal terminal device comprising:

a speech recognition unit for recognizing inputted speech;

an image recognition unit for recognizing an inputted image;

an instruction recognition unit for recognizing an inputted instruction;

neural network units respectively provided in at least two of said speech, image and instruction recognition units;

a bus unit operatively connected to said respective recognition units;

a processor unit operatively connected to said bus unit to perform processing upon the speech, image and instruction respectively recognized by said recognition units;

a memory unit operatively connected to said bus unit; and

a control unit for exercising control over information exchange between said respective recognition units and said memory unit under the control of said processor unit.

2. A customized personal terminal device according to claim 1, further comprising:

a sample data supplying unit for supplying sample data of either of speech and image to one of said speech recognition unit and said image recognition unit;

said neural network units of at least one of said speech recognition unit and said image recognition unit comprising a neural network for recognition;

said memory unit comprising a unit for storing output expectation of a neural network for recognition of one of said speech recognition unit and said image recognition unit corresponding to said sample data into said memory unit; and

said processor unit comprising a unit for reading and comparing the result of recognition of sample data in one of said speech recognition unit and said image recognition unit and said output expectation corresponding to said sample data stored in said memory unit to thereby derive an error, and a unit for adjusting weight factors of the neural network for recognition of one of said speech recognition unit and said image recognition unit so as to make said error less than a predetermined value.

3. A customised personal terminal device according to claim 2, comprising:

said memory unit comprising a unit for storing said sample data therein beforehand; and

said sample data supplying unit comprising a unit for reading out sample data stored in said memory unit and for supplying the sample data thus read out to one of said speech recognition unit and said image recognition unit as said sample data.

4. A customised personal terminal device according to claim 3, said sample data supplying unit further comprising:

a speech sensing unit for speech input;

an image sensing unit for image input; and

a unit for supplying either of speech and image sensed by one of said speech sensing unit for speech input and said image sensing unit for image input and either one of sample data read out from said memory unit selectively to one of said speech recognition unit and said image recognition unit.

5. A customised personal terminal device according to claim 1, comprising:

the neural network unit of said speech recognition unit comprising a neural network for speech synthesis; and

said speech recognition unit comprising speech synthesis unit having said neural network for speech synthesis for synthesizing and outputting speech on the basis of text data supplied from one of said processor unit and said memory unit via said bus unit, and an electro-acoustic transducer unit for outputting said synthesized speech as speech.

6. A customised personal terminal device according to claim 1, further comprising:

a speech detection unit for detecting input speech and supplying the input speech thus detected to said speech recognition unit;

a unit for inputting expectation of said input speech and supplying said expectation to said processor unit;

said speech recognition unit comprising a neural network for speech recognition; and

said processor unit comprising a unit for comparing the result of recognition of said input speech performed by said speech recognition unit with said expectation to derive an error and a unit for adjusting weight factors of said neural network for speech recognition so as to make said error less than a predetermined value.

7. A customised personal terminal device according to claim 1, further comprising:

an image detection unit for detecting an input image and supplying it to said image recognition unit;

a unit for inputting an expectation of said input image and supplying said expectation to said processor unit;

a display unit for displaying the result of recognition performed by said image recognition unit;

said image recognition unit comprising a neural network for image recognition; and

said processor unit comprising a unit for comparing the result of recognition of said input image performed by said image recognition unit with said expectation to derive an error and a unit for adjusting weight factors of said neural network for image recognition so as to make said error less than a predetermined value.

8. A customised personal terminal device according to claim 1, wherein

said processor unit comprises a unit for performing recognition processing of speech or image on the basis of results of recognition respectively performed by said speech recognition unit and said image recognition unit.

9. A customised personal terminal device according to claim 1, wherein said speech recognition unit comprises a unit for performing speech recognition processing on the basis of the result of recognition performed by said image recognition unit and the inputted speech.

10. A customised personal terminal device according to claim 1, wherein said image recognition unit comprises a unit for performing image recognition processing on the basis of the result of recognition performed by said speech recognition unit and the inputted image.

11. A customised personal terminal device according to claim 1, wherein said instruction recognition unit comprises:

a unit for registering a plurality of instructions beforehand;

a neural network for instruction recognition responsive to the result of recognition of the inputted speech or image performed by said speech recognition unit or said image recognition unit for selecting one instruction corresponding to said result of recognition out of a plurality of instructions stored in said registration unit; and

a unit for judging whether said inputted speech or image indicates a correct instruction or not on the basis of said selected instruction and said result of recognition.

12. A customised personal terminal device according to claim 11, wherein

said judgment unit comprises a unit for judging whether the result of recognition performed by said speech recognition unit or said image recognition unit coincides with said selected instruction or not and for outputting the result of judgment to said processor unit, and

said processor unit comprises a unit for performing processing in accordance with said selected instruction when the result of judgment formed by said judgment unit indicates coincidence.

13. A customised personal terminal device according to claim 12, wherein said terminal device further comprises a unit operatively connected to said bus unit to indicate an inputted signal, and said processor unit comprises a unit for supplying said selected instruction to said indication unit to cause indication when the result of judgment performed by said judgment unit indicates noncoincidence.

14. A customised personal terminal device according to claim 12, wherein said processor unit comprises a unit responsive to noncoincidence obtained as a result of judgment performed by said judgment unit for registering the result of recognition performed by said speech recognition unit or said image recognition unit into said registration unit in a manner that said recognition result is recognized as said selected instruction.

15. A customised personal terminal device according to claim 1, further comprising:

a memory unit so connected to said bus unit as to be mountable and detachable; and

said control unit comprising a unit for making possible data exchange between said respective recognition unit and said memory unit.

16. A customised personal terminal device according to claim 15, wherein

said control unit comprises a unit for making it possible to transmit/receive data in neutral network units included in said respective recognition unit to/from said memory unit.

17. A customised personal terminal device according to claim 16, wherein said control unit comprises a unit capable of down-loading weight factors in the neural network units included in said respective recognition unit to said memory unit and up-loading weight factors stored in said memory unit to said neural network unit.

18. A customised personal terminal device according to claim 15, further comprising a unit for inhibiting data to be written into said memory unit.

19. A customised personal terminal device according to claim 15, further comprising a unit for inhibiting data to be read out from said memory unit.

20. A customised personal terminal device according to claim 1, further comprising a unit for inhibiting data to be read out from at least one of said speech, image and instruction recognition units.

21. A customised personal terminal device according to claim 20, wherein said read inhibit unit comprises a unit for inhibiting data in the neural network units included in said speech, image and instruction recognition units to be read out.

22. A customised personal terminal device according to claim 1, wherein at least one of said image recognition unit and said speech recognition unit comprises a mount/detach unit for making it mounted to/detached from said bus unit.

23. A customised personal terminal device according to claim 22, wherein said recognition unit comprising said mount/detach unit has a memory unit for storing the result of recognition.

24. A customised personal terminal device according to claim 23, wherein said memory unit comprises a unit for making it mounted to/detached from said recognition unit comprising said mount/detach unit or said bus unit.

25. A customised personal terminal device according to claim 1, wherein at least one of said speech, image and instruction recognition units comprises a fuzzy control unit, and

said control unit comprises a unit for making possible data exchange between said fuzzy control unit and said memory unit.

26. A customised personal terminal device according to claim 25, further comprising:

a memory unit so connected to said bus unit as to be mountable and detachable; and

said control unit comprising a unit for making possible data exchange between said fuzzy control unit and said memory unit.

27. A customised personal terminal device according to claim 1, wherein at least one of said speech, image and instruction recognition units comprises a knowledge base unit, and

said control unit comprises a unit for making possible data exchange between said knowledge base unit and said memory unit.

28. A customised personal terminal device according to claim 27, further comprising:

a memory unit so connected to said bus unit as to be mountable and detachable; and

said control unit comprising a unit for making possible data exchange between said knowledge base unit and said memory unit.

29. A customised personal terminal device according to claim 1, wherein said speech recognition unit comprises a unit for selectively inputting:

input speech of the operator of said terminal device;

speech from another terminal device operatively connected to said bus unit; and

speech read out from said memory unit.

30. A customised personal terminal device according to claim 1, wherein said processor unit comprises a host processor and a peripheral processor, and said peripheral processor comprises a unit for making it mounted to/detached from said bus unit.

31. A customised personal terminal device comprising:

a first recognition unit for recognizing inputted speech or image, said first recognition unit having a neural network unit;

a second recognition unit for recognizing an inputted instruction, said second recognition unit having a neural network unit;

a bus unit operatively connected to said respective recognition units;

a processor unit operatively connected to said bus unit to perform processing upon either speech or image and upon instruction, said speech, image and instruction being respectively recognized by said recognition units;

a memory unit operatively connected to said bus unit; and

a control unit for exercising control over information exchange between said respective recognition units and said memory unit under the control of said processor unit.