JPS63129765A - Telephone set with voice recognition function - Google Patents

Abstract

PURPOSE:To use a specific call recognition system with good recognition rate even in a telephone set used by many and unspecified number of users such as a public telephone set by providing a voice recognition section inputting/ outputting a reference voice pattern data to a storage card medium. CONSTITUTION:A signal processing section 15 of a voice recognition section 14 extracts an input voice pattern data 117 suitable for the voice recognition from a transmission/reception voice signal 105 inputted from a signal circuit 8, the pattern data 117 is stored in an input buffer 5 and used for the collation with the reference voice pattern data 112. The basic voice pattern data 112 is supplied to a collation arithmetic section 12 from a portable storage card medium 7 through a basic voice pattern data input section 6, and the result of collation by the collation arithmetic section 12 is fed to a decision section 11 as a similarity parameter 111. A control section 10 applies generation of, e.g., a dial impulse 114 and the transmission/reception of a command with the device connected to the subscriber line 116 corresponding to the result 110 of recognition.

Description

[Detailed description of the invention] [Industrial application field] The present invention relates to a telephone having a voice recognition function.

[Conventional technology]

Conventionally, this type of telephone with voice recognition function has a specific speaker recognition method and a non-specific speaker recognition method.
Both systems have voice pattern data, which serves as a reference for verification, as registered data in memory inside and outside the phone.

[Problem that the invention seeks to solve]

In the above-mentioned conventional telephone with specific speaker/speech recognition function,
Because the voice pattern data of a specific speaker, which serves as a reference for verification, is stored as registered data in the telephone or in a device attached to the telephone network, there is a limit to the number of people who can register due to physical memory constraints, It is not suitable for telephones used by an unspecified number of users. On the other hand, it is possible to use a speaker-independent recognition method, but it has the disadvantage that the number of recognized words cannot be increased much due to its poor recognition rate.

[Means for solving problems]

A telephone with a voice recognition function according to the present invention includes a handset that converts an input voice signal into a transmission voice signal, a receiver that converts a reception and transmission voice signal into an output voice signal, and a telephone receiver that inputs the transmission voice signal and transmits and receives a transmission voice signal. , a signal circuit that outputs the received and transmitted audio signals and transmits and receives the received and transmitted audio signals, a dial that outputs dial number information, and a dial impulse generator that inputs the dial number information and other dial number information and outputs a dial impulse. A circuit, a man-machine interface that inputs and outputs the man-machine interface,
a control section that inputs the recognition result of the man-machine interface and outputs dial number information and control signals of the side; and a control section that inputs the received and transmitted voice signals, dial pulses, and control signals and transmits and receives them to and from the subscriber line. A communication circuit, a signal processing unit that converts the input buffer into a man-machine interface, an input buffer that inputs the man-machine interface and outputs transmitted and received voice data, and inputs and outputs reference voice pattern data to and from a card-shaped storage medium. a speech recognition section having a reference speech pattern data input section that matches the transmitted and received speech data with the reference speech pattern data and outputs a similarity parameter; and a judgment section that obtains a recognition result from the similarity parameter. It is characterized by comprising:

〔Example〕

Next, the present invention will be explained with reference to the drawings.

FIG. 1 shows an embodiment of the invention. The voice signal 101 input to the transmitter 1 is transmitted to the subscriber line 116 through the signal circuit 81 and the speech circuit 13. Conversely, subscriber line 1
The received and transmitted audio signals inputted from 16 are sent to the communication circuit 13.
Signal circuit 8. The signal passes through the receiver 2 and becomes an output audio signal 102. The dial number information 106 input from the dial 3 is transmitted to the subscriber line 116 as impulse information through the dial impulse generation circuit 91 and the telephone call circuit 13. The man-machine interface 4 refers to an indicator etc. for informing the telephone user of the internal state of the telephone. Note that 103 is a transmitting audio signal, 104 is a receiving audio signal, 113 is a receiving/transmitting audio signal, 115 is a control signal, and 107 is a man-machine interface signal.

Next, the speech recognition section 14 will be explained. In the signal processing unit 15, the transmission/reception audio signal 105 inputted from the signal circuit 8 is processed.
Input voice pattern data 117 (digital value) suitable for voice recognition is extracted from (analog value). This pattern data 117 is stored in the input buffer 5 and used for comparison calculation with the reference voice pattern data 112. In addition, 10
8 is transmitted/received audio data. Basic audio pattern data 1
12 is supplied from the portable card-shaped storage medium 7 to the verification calculation section 12 through the reference voice pattern data input section 6.

The result of the matching calculation in the matching calculation unit 12 is sent to the determining unit 11 as a similarity parameter 111. The determination unit 11 obtains a recognition result 110 based on the similarity parameter 111 and sends it to the control unit 10 . The control unit 1o performs a predefined operation corresponding to the recognition result 110. Various predefined operations can be considered, including, for example, generating dial impulses 114, controlling man-machine interface 4, and sending and receiving commands to and from equipment connected to subscriber line 116.

〔Effect of the invention〕

As explained above, the present invention does not require pre-registration of voice pattern data, which serves as a reference for recognition verification, in the telephone or in a device attached to the telephone network. This has the advantage that a specific speaker recognition method with a high recognition rate can be used even in telephones used by an unspecified number of users.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an embodiment of a telephone with voice recognition function of the present invention. 1...Telephone, 2...Receiver, 3...Dial,
4...Man-machine interface, 5...Input buffer, 6...Reference audio pattern data input section, 7
...Portable card-like storage medium, 8...Signal circuit, 9
...Dial pulse generation circuit, 10...Control section, 11...Judgment section, 12...Verification calculation section, 13
...Telephone circuit, 14...Speech recognition unit, 15...Signal processing unit, 101...Input audio signal, 102...Output audio signal, 10B...Transmission audio signal, 104...
Receiving/transmitting audio signal, 105... Transmitting/receiving audio signal, 10
6...Dial number information, 107...Man-machine interface (H'Ff, 108...Transmitted/received voice data, 109...Dial number information, 110...
・Recognition result, 111...similarity parameter, 112・
...Reference audio pattern data, 113...Receiving and transmitting audio signal, 114...Dial impulse, 115...Control signal, 116...Subscriber line, 117...
- Input audio pattern data.

Claims (1)

[Claims] A transmitter that converts an input audio signal into a transmitted audio signal, a handset that converts a received audio signal into an output audio signal, a transmitted/received audio signal that inputs the transmitted audio signal, and a received/transmitted audio signal that outputs the received audio signal. a signal circuit for transmitting and receiving information, a dial for outputting dial number information, a dial impulse generating circuit for inputting the dial number information and other dial number information and outputting a dial impulse, and a man-machine interface for inputting and outputting the dial number information. - A machine interface, a control unit that inputs the recognition result of the man-machine interface and outputs the separate dial number information and a control signal, and a control unit that inputs the received and transmitted voice signals, dial pulses, and control signals and that receives the subscriber a speech circuit that transmits and receives data to and from the line, and a signal processing unit that converts the input buffer into a man-machine interface;
an input buffer for inputting the man-machine interface and outputting transmitted/received audio data; a reference audio pattern data input section for inputting/outputting reference audio pattern data to/from the card-shaped storage medium; and comparing the transmitted/received audio data with the reference audio pattern data. 1. A telephone with a voice recognition function, comprising: a verification calculation section that outputs a similarity parameter; and a voice recognition section that has a determination section that obtains a recognition result from the similarity parameter.