JPS59176836A - Processing system for sound input data - Google Patents

Abstract

PURPOSE:To transmit the result of recognition to a data processing part different in type by transmitting the reception style of a data processing part to a sound input part by the data processing part itself when said data processing part is actuated and therefore deciding the transmission style of a sound recognizing part. CONSTITUTION:A personal computer 2 contains a type information memory region M where the type information (m) to be decoded by a decoder 27 is stored in a memory 25. Then the information (m) is transmitted to a transmitting circuit 16 of a sound recognizing part 1 from a central processing part 21 when the computer 2 is actuated. The part 1 changes the sound signals sent from a microphone 11 into a pattern through a patterning circuit 12. This pattern is stored temporarily in an input sound pattern memory 13, and a pattern recognizing circuit 15 decides a sound pattern in a registered sound pattern memory 14 to which the pattern produced from the circuit 15 is corresponding. Then the register number of the decided sound pattern is delivered to the output circuit 16. The circuit 16 decides the transmission style from the information (m) given from the computer 2 and transmits it to the computer 2.

Description

[Detailed description of the invention]

(a) Industrial Application Field The present invention relates to a voice input data processing system that combines a voice recognition section and a data processing section. (b) The conventional voice input data processing system is a personal computer (hereinafter referred to as a personal computer) for performing accounting processing, etc.7, for example.
A word processor (:) that recognizes speech and transmits an output signal corresponding to the speech is combined with a speech recognition section that

[Second step is to operate these data processing units. In such a conventional system, the recognition result of the voice recognition unit is transmitted to the data processing unit in the form of a command code or data code expressed in ASCII code, for example. The transmission format of the recognition results is fixed, and there is an inconvenience that the recognition results are not received by data processing units with different formats. That is, for example, as a code indicating the end of transmission, the BAsIC three words “
Form that only uses OR” (carriage return) (7)
/(Sokon and its code “OR@(!:”LF
- (Line field) Since the computer of the type used in combination coexists, the data processing 3 from the voice recognition section
Even if an attempt was made to transmit the recognition results to the personal computer serving as the ¥ section, there were cases in which the recognition results could only be transmitted to the personal computer (2) due to the difference in format. (c) Purpose of the Invention The present invention has been made in view of the above points, and is a speech recognition unit capable of transmitting recognition results to any data processing unit C2 having a different code system format! The present invention provides a voice input data processing system with the following features. B) Structure of the Invention The voice input data processing system of the present invention is a system in which a voice recognition section and a data processing section are disconnected.
A transmitting circuit is provided to convert the recognition result of two-person speech into a code system in a format that can be received by the data processing unit, and a storage means is provided that stores format information that instructs the data processing unit to format the code system visually. When the data processing section is activated (2), by transmitting the format information from the storage means to the transmission circuit of the speech recognition section, the transmission format from the transmission circuit is determined relative to the reception format at the data processing section. ζ) Embodiment FIG. 1 shows the configuration of an embodiment of the voice input data processing system of the present invention. The system in the figure shows an example of RAS I (a personal computer (2)?I' that uses three languages) as a data processing unit, and first we will describe the configuration of the speech recognition unit (1) that is connected to this personal computer (2). In the voice recognition unit (1), aD is a microphone that converts voice into an electrical voice signal ζ, and L121 is a patterning circuit that patterns the voice signal from the microphone CIl+, for example, from frequency spectrum parameters. A voice pattern is obtained. (131 is an input voice pattern memory that temporarily stores voice patterns from the patterning circuit 2, and 141 is a registered voice pattern memory that stores a plurality of registered voice patterns in advance.) .Scream is the registered back door pattern memo!
This is a pattern recognition circuit that recognizes one input voice pattern t pattern, and outputs a recognition result indicating the voice of the registered voice pattern with the minimum error from the input voice pattern, for example, with its registration number. The wholesaler is an output circuit for outputting the recognition result indicated by the number from the pattern recognition circuit as a data code expressed, for example, in ASCII code, and converts this data code into a format that can be received by the personal computer (2). It has a function to add either "OR" or "(3R"=LF" command code) indicating that the output of the result has been completed after the data code indicating the recognition result. This “OR” or “OR”
The selection of “LF” is based on the format information obtained from the computer (2).
I will do it depending on the situation. On the other hand, in the personal computer (2), (2il is the central processing circuit, Keyho F (22+, ORT (33, 7)
-! I'/9Q4J, has a basic configuration with a memory circuit (c) connected to it, and also has the above speech recognition section (1)
The recognition result from external human power (for example, or for that human power) is stored in the paper memory t20 and then after that data ('OR' indicating the end followed by 'OR', or 'OR' or 'LF').
Decoder C17 for the command code of one of the 'age'
), the contents of this buffer memory CQ can only be introduced into the central processing circuit I21) after being decoded. In addition, the above memory circuit (2!9(”:) is the above decoder (2!9).
7) stores in advance the format information m of the received format from the eyes indicating whether to decode the command code °OR" or -OR"LP" by inputting it from the keyboard 12]). A format information storage area [Ml is provided, and the format information m stored in this area (MJ) is stored in the central processing unit CI!1 when the personal computer (2) is started.
1 to the transmission circuit αe (-) of the voice recognition center (1). In the voice input data processing system configured as described above, when the personal computer (2) is started, ) Depending on the format information storage area + format information m of Ml, the transmission format in the transmission circuit aI2 of the speech recognition unit (1) is a command code of 'OR' added after the data code indicating the recognition result. , the father either adds the command code of °OR""LF", or M of the decoder ρη of the PC (2)
, the power will be installed in a format compatible with the image interpretation method. Next, the specific configuration of the above-mentioned transmission circuit (6 circuits (16i) is shown in FIG. 2nd and 6th register 0
11. C(3,ai''4r is a three-stage shift register connected in series, and this second register t32JI- is a shift register in which the recognition result from the pattern recognition circuit is converted into ASCII code by the converter (4) and introduced. On the other hand, the first command register (51) in which the command code of "C](" is stored is stored in the first register C3i1.
and the second command register (52) in which the command code of OR “LP” is stored is the first command register from the personal computer (2).
The opening/closing control is controlled by the format information m1g2, which is connected via goo)gl, and goo)g2Y, and further through goo)g3'&. Therefore, when the personal computer (2) is started, if "L0 level format information m1 is obtained", the gate g1 is opened and the "OR" command data of the first command register (51) is is introduced into the first register cll via gate g1 and gate g5 (open state).On the other hand, if "H" level format information m1 is obtained from the personal computer (2), By forming the gate g2vM, the second command register (52) =CH LF-
The command data is sent to the first register cll) (through Goo)g2'P, and then to the open state through Goo)g3Y.
Two are introduced. Therefore, the shift register (3) outputs one of the ASCII code strings selected from, for example, "6"OR" or -5"-OR@"LF", and the personal computer (
2) (: Transmitted. Also, in this transmitting circuit a6J, there are dozens of recognition results from the pattern recognition circuit, and these recognition results are 1
In the case of C:, which is represented by a digit number and a two-digit number, it also has a function to arrange a one-digit number, for example "3", into a two-digit format such as "03." The personal computer (2) has a counter Z that detects the total number of bytes of the code string stored in the decoder (instead of 271) of the code string stored in the C buffer memory (to). 1: This corresponds to the case where the contents of this buffer memory (to) are introduced into the central processing circuit. Such a specific configuration is shown in Figure 2 (2). As shown, the third register of the shift register (3) stores the data code of "0° (zero)" in the data register (6) lag-)g4,
This gate g4 is opened and closed by the digit signal K indicating the number of digits of the number code obtained from the converter (4), while the gate g5 is connected to the second gate from the personal computer (2). It is opened and closed using the format information m2. So, computer (2)'? When ff is activated, if "H" level format information m2 is obtained, this information m2 opens gate g5 and closes gate g3 to perform the above-mentioned -OR",=CH"'LF". A command code becomes unnecessary.In this state, when a digit signal K indicating a one-digit number data code is obtained from the converter (4), data) g4 is opened and the sixth register (c) is opened. (A data code of “0.” of the data register (6) is introduced. Therefore, from the shift register (3), for example, aO°
An ASCII code string of ``3 degrees is output, and the computer (2
). In addition, even if the data register (6) stores the ASCII code of C2 "SP・(space) instead of the ASCII code of "0''', for example, "6・ can be represented with "mouth 3・". Therefore, it is processed in the same way as the two-digit format described above.
The registration number of the registered voice pattern that is most similar to the input voice pattern is used as the recognition result transmitted from the input voice pattern to the PC (2). It may be configured to transmit data, and in this case, the first. By transmitting the second and $3 registration numbers together with the degree of similarity l, even if the first registration number is misrecognized, the computer (
) side (input C''' at the time of second lever, appropriate input Z can be reselected from among these.Furthermore, the voice recognition section (
Each registered crying pattern in the registered voice pattern memory I in 1) and the input voice pattern in the input voice pattern memory 03 can be transmitted as they are to the personal computer (2) and displayed on, for example, 0RTq, or can be saved in the memory circuit (2 chants: In other words, the above-mentioned personal computer (
2) Using v (2) In addition to data processing such as the original accounting processing of this computer (2), processing to modify voice patterns to improve the recognition rate in the voice recognition unit (1), etc. As a result, the data processing section of the present invention can be executed more reliably than the original data processing section.However, the data processing section of the present invention is not limited to a personal computer (2), but can be used as a data processing section of any reception format. In order to adapt the transmission format of the speech recognition unit (1) to the data processing unit, it is sufficient if the format information can be stored in the data processing unit, such as a word processor. Effects In the voice input data processing system of the present invention, when the data processing unit is activated, the data processing unit transmits format information indicating the reception format Z from the eyes to the voice input unit, and this (depending on the transmission format of the voice recognition unit) Since it is compatible with the reception format of the data processing unit, it is possible to transmit the recognition results from the speech recognition unit to any data processing unit with a different code system format, and it is possible to change the transmission format of the recognition gIS. Therefore, the versatility of the speech recognition unit itself is improved, and the configuration of this type of system can be expanded to a large extent. 44. Brief Description of Drawings FIG. 2 is a block diagram showing the configuration 7 of an embodiment of the input data processing system, and FIG. 2 is a block circuit diagram of the main parts of the voice input data processing system of the present invention. (1)...Speech recognition section, (2) ...PC, oto microphone, a7J...patterned circuit, a4
Record 1...Voice pattern memo'J, [1ω...pattern recognition circuit, cIe...transmission circuit, circle...central processing circuit, zero...memory circuit, (c)... - Backup memory, 21... Decoder, [Ml... Format information storage area.

Claims (1)

[Claims] 1) Voice input data processing coupled with a voice recognition unit that recognizes human voice and a data processing unit that performs data processing based on a command code or data code in a predetermined format. In the system, the voice recognition unit includes a microphone that converts voice into an electrical voice signal, a patterning circuit that converts the voice signal from the microphone into patterns, and a plurality of registered voice patterns that are stored in advance. a registered back door pattern memory; a pattern recognition circuit that recognizes the input voice pattern obtained from the patterning circuit based on a plurality of registered voice patterns of the registered back door pattern memory; and a recognition result obtained from the pattern recognition circuit. a transmitting circuit that converts the received type I code system in the data processing section and transmits it to the data processing section;
A storage means is provided for storing format information for instructing the format of the code system, and the format information is transmitted from the storage means to the transmission circuit of the speech recognition section when the data processing section is activated. A voice input data processing system characterized in that a transmission format Z from a transmission circuit is adapted to a reception format at a data processing section.