JPH04163618A - Sound operation computer - Google Patents

Abstract

PURPOSE:To easily input a command even if the performance of a sound recognition device is low by limiting sound being the object of recognition to sound related to an object being the object of a present operation at the time of executing a computer operation in the sound recognition device. CONSTITUTION:Control information 107 of the object and sound information 118 for improving the performance of sound recognition in a data structure itself are added to a secondary storage device 102 and it stores them. When plural objects being the object of the present operation are decided, sound information peculiar to the object is automatically transmitted to the sound recognition device 101 and it is operated. It restricts the reading, the feature amount and the speaker of the object of recognition based on sound information peculiar to the object and recognizes sound. Thus, the command can easily be inputted even if the sound recognition device has low performance.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a computer equipped with a speech recognition device.

(Prior Art) Conventionally, the method of operating a computer has been as follows.

The computer, approximately as shown in Figure 2 (A),
Computer (keyboard 205 for inputting commands, control device 2O6 for executing programs, main memory 2 for storing programs being executed and data being processed)
OB is equipped with a secondary storage device [202] for storing executable programs and a large amount of data that does not need to be processed at high speed, and is not equipped with a voice recognition device for inputting commands. Here, illustrations and explanations of other output devices, Tsushin M, etc. will be omitted.

A program called an operating system exists inside the main storage device 208, and defines the operating procedure of the control device 206.

The secondary memory M2O2 has a storage area called TiLeft 1 area 203, which contains management information 2 such as file names (hereinafter referred to as file names), attributes, and storage locations of a large number of files.
07 is saved. Here, the following explanation will be given assuming that the file names are ``Word Processor'' for a word processor program file that has an executable file attribute, and ``Spreadsheet'' for a spreadsheet program file. .

First, the procedure for operating the computer using the keyboard 205 will be explained with reference to FIGS. 2(B) and 2(C). The keyboard 205 allows the user to select, for example, "spreadsheet".
When a command is input, the control table M2O6, based on the program of the operating system 209, stores the management information with the file name corresponding to this command name in the directory area 203 based on the directory specification (250) and bus specification (251). Search inside the file and read the attribute corresponding to the file name (shown by the dotted line in Figure 2 (B)). If the file with this file name has the attribute of being an executable file, the file 204 (in this case, a spreadsheet program file) is read based on the storage location (not shown) in the management information 207, The data is transferred to the main memory N2O3 and the program in this file 204 starts running (as shown by the dotted line in FIG. 2(C)).

Normally, the TiLeft area mentioned above can refer to other directory areas, and the TiLeft area to be used can also be specified by a command from the keyboard (this specification can be done using Directory Specification 2).
50). Further, normally, the directory area to be searched for in the file search can be arbitrarily designated (this designation is referred to as bus designation 251).

On the other hand, commands called internal commands are processed by the operating system 209 and do not necessarily require reading from the secondary storage device 202 as described above. In the case of this internal command, the execution procedure generally exists in the main memory 208 or the control table 206, and if the input command is an internal command, the control table M2O6 immediately executes its execution.

(Problems to be Solved by the Invention) However, in the conventional operating method described above, since the keyboard is the only means for inputting commands, keyboard input requires effort and time, and it takes time to complete the operation.

In addition, in computers with conventional configurations, in order to specify and operate files by voice recognition input instead of keyboard input, it is necessary to use a large language service that can handle all command names and word searches for the number of files. Therefore, a high-performance speech recognition device was needed.

An object of the present invention is to eliminate the above-mentioned problem that it takes time and effort to operate through keyboard input, and to provide a computer that can be operated by voice commands using a voice recognition device that is compatible with a small number of language languages. It is in.

(Means for Solving the Problem) In order to achieve this object, the voice-operated computer of the present invention includes a keyboard for inputting commands, a control device connected to the keyboard for executing a program, and a control device for executing a program. The main memory device is connected to the controller and stores the programs being executed and the data being processed, and the main memory device is connected to the control device to store programs that can be easily processed and data that does not need to be processed at high speed.
The secondary storage device is equipped with a voice recognition device that is connected to the control device and performs speech recognition processing and outputs the recognition result to the control device.The secondary storage device stores object management information and its data structure itself. (Speech information is added and stored in order to improve the performance of speech recognition, and the control device receives speech information specific to the object currently being operated in response to commands entered from the keyboard.) , is characterized in that it has a function of transmitting it to a speech recognition device as a reference pattern for recognition.

(Function) Computers usually do not operate on all executable commands and programs, but specify search areas for commands and programs called a search bus, or display the current directory to the user in a fixed manner. , has a means to limit the data, programs, etc. that are subject to operations.

In this invention, the management information and data structure of objects (hereinafter referred to as objects) such as data, programs, pictograms (icons), buttons, etc. that are the targets of current operations for the user, and the data structure itself, are unique to each object. , information to improve the performance of speech recognition, such as pronunciation, voice features, speaker, etc. (hereinafter referred to as speech information) was added, and multiple objects to be the target of the current operation were determined. Sometimes, object-specific voice information is automatically transmitted to the voice recognition device, and the voice recognition bag N is operated. The speech recognition device uses this object-specific speech information to determine the reading/toku8! of the object to be recognized. Speech recognition is performed by limiting the amount and speakers. When a voice is uttered and a recognition result is obtained, the recognition result is input to the control device, and the computer is operated by processing similar to command input using a keyboard or mouse.

(Embodiments) Hereinafter, embodiments of the present invention will be described with reference to the drawings.

An embodiment of the present invention will be described with reference to FIGS. 1, 3, and 4.

The overall configuration of the voice-operated computer of the present invention is approximately as shown in FIG. 1(A). This computer has a keyboard 1 for inputting commands.
05. A control device 106 that executes programs, and a main storage device 10 that stores programs that are being executed and data that is being processed.
8. Secondary storage for storing executable programs and large amounts of data that does not need to be processed at high speed! 102! Be prepared. Here, since other output devices and communication devices are not directly related to the present invention, illustrations and descriptions thereof will be omitted. The secondary storage device M102 usually has a plurality of storage areas called directory areas 103. Up to this point, the configuration is similar to that of a normal computer.

In this invention, a voice recognition device 701 is provided in addition to a keyboard] 05 as an input device. In this embodiment, the speech recognition device 1101 is assumed to be a word-by-word speech recognition device, for example. In addition, in the present invention, the management information 107 such as file name, attribute, storage location, etc. includes information 118 specifically for improving speech recognition performance.
is saved.

In addition, FIG. 3 is a configuration diagram showing details of the control table M106,
FIG. 4 shows a flowchart showing the operation of the control device 708.

In FIG. 3, 301 is an internal command storage means;
Reference numeral 2 indicates a command searching means, and 303 indicates a command execution means. The internal command storage means 307 stores internal command names, beauty procedures, etc., as well as voice characteristic patterns unique to the internal command, and the voice characteristic patterns are transferred to the voice recognition device 01 as necessary. Command search means 3
02 searches for the input command 8.

The command search means 302 is configured to be able to exchange information with the secondary storage device M]02 in order to search for external commands. Command execution means 303 executes commands. Therefore, this command execution means 303 is configured to be able to exchange information with the keyboard 105 and the voice recognition device 101, and the secondary storage device 1
02 and the main storage device 108.

In FIG. 4, (S]) to (Sl○) are the control device 106.
shows the internal processing steps.

In the example described here, it is assumed that the name of one directory area 103 is [Business J].
For a word processor program file with the file name "Word Processor", the word audio "Word Processor J" is displayed.
The sound characteristic button of the word "spreadsheet j" is saved in the management information 107, and for the spreadsheet program file with the file name "spreadsheet", the sound characteristic button of the word "spreadsheet j" is saved in the management information 107. I'm going to leave it.

Further, it is assumed that the internal command storage means 301 in FIG. 3 stores in advance the names specific to the internal commands of the operating system and the voice characteristic buttons corresponding to their operations.

Next, an example of the operation of the voice-operated computer in an embodiment under such circumstances is shown in FIGS.
), and will be explained with reference to Figures 3 and 4.

<Deleting/Rewriting Audio Feature Buttons> First, the bus designation and directory designation are set and changed by inputting the action command when starting up the computer. When setting or changing this bus designation or directory designation, the control device 106 first erases the audio characteristic button corresponding to the bus or directory that was used immediately before responding to the bus designation or directory designation after the designation or change. Then, the audio feature button of the file that may be the target of the current operation is sent to the speech recognition device 101.

In this case, for example, MS
- In the case of an operating system having commands similar to DO3, the operator normally inputs a command such as "cd\business" or a command for changing the directory designation using the keyboard. At this time, first, the control device 10
This command line is input through the command retrieval means 302 inside the computer 6 or through the processing steps S1 and S21 in FIG.

In processing step S3, the command search means 302 performs a comparison with a character string unique to an internal command stored in advance in the internal command storage means 301 as an internal command of the operating system. Since this command "cd" is a 1fi class of internal commands, the comparison results in a match with the internal command, and based on the determination in S4, the process proceeds to step S5. If the input command is an external command or a program, processing step S1
0, the external command or program is executed and the process returns to step S2.

In processing step S5, the command search means 302 determines whether this internal command input from the keyboard 105 is a change in directory designation or bus designation. Since this command "ccj" is a command for specifying a directory, the process proceeds to step 86.6 If the input command does not change the directory specification or bus specification, the internal command is executed in process step S9, and the process proceeds to process step S2. return.

In processing step S6, the command execution means 303 instructs the voice recognition device]01 to erase the voice characteristic button inside the voice recognition device 101.

Next, in processing step S7, the bus designated area 151 and the directory designated area 11 inside the main memory 108 are
0 is changed according to the contents of the input command. Here, the directory specification is the directory name′°\business”
Rewrite with .

Next, in step S8, voice characteristic buttons are transferred to the voice recognition device 1101 from the directory area of internal commands, external commands, and programs to be voice recognized by bus designation and directory designation. Therefore, here, first, the voice characteristic button corresponding to the internal command stored in the internal command storage means 301 is transferred to the voice recognition device 101. Next, the directory area t1i of external commands and programs existing in the directory "\business" existing in the secondary storage device N102 according to the information %9 of the current directory "\business" after the Tileft reference designation has been changed. 103, the corresponding voice feature buttons are transferred to the voice recognition device flO1. Here, the speech feature buttons "Word Processor" and "Spreadsheet" 118 stored in the directory area 103 of the programs "Word Processor" and "Spreadsheet" 104 existing in the directory "\Business" are transferred to the speech recognition device 101. be done. Furthermore, if necessary, the voice feature button is transferred from the external command and program directory area in the directory indicated by the current bus designation to the voice recognition device 101.

With the above steps, all the voice characteristic buttons corresponding to the object currently being operated are transferred to the voice recognition device 101.
Since preparations for inputting the next operation by voice recognition are complete, the process returns to step S2 to continue inputting and processing the next command.

In this way, in conjunction with the directory designation change command, the voices to be recognized by the voice recognition device 101 are r word processor J and r spreadsheet J (Fig. 1).

(B)). In this embodiment, the operation when the directory designation is changed has been described, and the same operation is performed when the bus designation is changed.

In this way, the standard button of the speech recognition device 101 is automatically rewritten with the information (speech information) that is specific to the object that is the object of the current operation. The voice to be inputted through a voice input means and the voice to be recognized for operating the computer using this input voice is limited to the voice related to the object currently being operated by the computer. Become.

<Operation by Speech Recognition Device> Next, the procedure for operating the computer by the speech recognition device 101 will be explained with reference to FIGS. 1(C) and (D). When the user utters a voice corresponding to a command, for example r spreadsheet J, to the voice recognition device tlO1, the voice recognition device 101 performs a conventional recognition process and outputs the recognition result “spreadsheet J” to the control device 106. The control device 106 searches the directory area 103 for the program file corresponding to this recognition result according to the processing flow (32 to S4.510) in FIG. 4 (indicated by a dotted line in FIG. 1C), The spreadsheet file 104 is read out based on the storage location in the management information 107, transferred to the main storage device M108, and execution of the program in this file 104 is started (as shown by the dotted line in FIG. 1(D)). ).Furthermore, if the uttered content is not a voice corresponding to the program, but corresponds to an internal command or a tato part command, the desired action corresponding to the uttered content is performed.

As in the embodiment described above, commands existing in the secondary storage device 102 (usually called external commands)
In addition to this, it is preferable to configure the system so that the execution of internal commands can also be operated by voice recognition. In this case,
The sound characteristic button corresponding to the internal command exists together with the management information (not shown) of the operating system file existing in the secondary storage device 102, and when the computer is powered on, the operating system file (
It is preferable that the voice feature button is transferred to the voice recognition device 101 when the command (not shown) is transferred from the secondary storage device 102 to the internal command storage means 301 in the control device.

In addition, in the same way that voice recognition is used to specify commands to the operating system, when voice recognition is used to specify commands and files to application programs, voice characteristics and application program file management areas are subject to manipulation. It is preferable to add this to the management information of the file.

Furthermore, in the above-mentioned embodiment, the word speech recognition bag N was assumed as the speech recognition device, but the speech recognition device in this invention does not necessarily have to be a word-by-word speech recognition device, but may be a continuous speech recognition device. There may be. In this case,
Instead of the sound feature button, it is preferable to add information such as file-specific pronunciation to the management information as sound-related information.

In general, the objects that are the targets of computer operations (hereinafter referred to as objects) are not only files on the secondary storage device, external commands, and internal commands inside the control device, but also objects on the C/T display. Data structures representing cursors, characters, icons, buttons, etc. can also be considered. In the embodiments described above, internal commands, external command files, and program files were cited as examples of objects, but objects are not necessarily limited to files. Data structures representing icons and buttons may be handled as objects.

(Effects of the Invention) According to the present invention, when an operation is performed on a computer using a speech recognition device, the speech to be recognized is limited to the speech related to the object (file, data structure) that is the object of the current operation. This allows commands to be input easily even when voice recognition performance is low and only a few words can be recognized.

[Brief explanation of the drawing]

Figures 1 (A) to (D) are diagrams for explaining the voice sweeping computer of the present invention, Figures 2 (A) to (C) are diagrams for explaining a conventional computer, and Figure 3 is a diagram for explaining the conventional computer. FIG. 4 is a diagram illustrating the control device included in the voice-operated computer of the present invention. FIG. 4 is an operational flowchart of FIG. 3. ]01...Speech recognition device, 102...Secondary storage device 103-...Directory area]04...File, 105...Keyboard'
106...Control device, 108...Main memory]09...Operating system]]○...Directory specification]51...Path specification.

Claims (1)

[Claims] (1) A keyboard for inputting commands, a control device connected to this keyboard to execute programs, a main storage device connected to this control device for storing programs being executed and data being processed, and the control device a secondary storage device that is connected to store executable programs and data that does not need to be processed at high speed, and a speech recognition device that is connected to the control device and performs recognition processing on vocalized sounds and outputs recognition results to the control device. The secondary storage device stores object management information and the data structure itself with voice information added to improve voice recognition performance, and the control device stores information from the keyboard. A voice-operated computer, characterized in that it has a function of transmitting voice information specific to the object currently being operated, as a reference pattern for recognition, to the voice recognition device according to an input command. .