JPH0540599A - Voice input device - Google Patents

Abstract

PURPOSE:To obtain the voice input device to be easily used equally to a 'Japanese syllabary (KANA)/Chinese character (KANJI) converting Japanese input' without requiring the change of an application software. CONSTITUTION:This device is composed of a data input means 1, data output means 2, voice input selection detecting means 3, voice input means 4, voice feature extracting means 5, data storing means 6, voice feature pattern matching means 7 and data input/output control means 8. Thus, without requiring the correction of the existent application software, the device can be used similarly to the normal 'KANA/KANJI converting Japanese input'.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice input device.

[0002]

2. Description of the Related Art A voice input device is considered to be a very efficient information input means. However, the voice input device has not yet become a general information input means. There can be several reasons.

Of course, the first thing to think of is,
This means that when using the voice input device, it is necessary to prepare hardware including it and application software capable of handling voice input in advance.
A very large amount of investment in hardware is required to newly install these.

An even greater problem is the definitive lack of application software that can handle voice recognition. In the conventional technology, when trying to handle voice recognition, it was necessary for the application software to handle it. However, creating software requires significantly more human investment than creating hardware. Therefore, it is practically impossible to make existing software compatible with voice recognition.

Further, even if the conventional software is rewritten into one that is compatible with voice recognition input, if the usage method of the software is changed by that, it will be a major factor rejected by the user.

In addition to the above-mentioned problems, a problem of the voice recognition technology itself is that it is difficult to cut out the utterance section even in the noise of the normal office environment. Further, since the voice recognition processing requires a considerable amount of computer resources, keeping it running may adversely affect other processing.

As described above, the conventional technique has a large number of problems in terms of hardware and software as described above, and even if all of these problems are solved, further problems still remain. It was very difficult to prepare an environment that is easier to use.

[0008]

The problem to be solved by the present invention is as described above, and the object of the present invention is to minimize the existing hardware to solve the above problems. By adding the hardware for voice input, it can be used without changing the existing application software, and by realizing a voice input device that allows the user to easily select voice input and key input. is there.

[0009]

FIG. 1 is a schematic view of the concept of the present invention. The present invention will be described with reference to FIG. 1. The present invention includes: 1) one or more sets of data input means 1; 2) one or more sets of data output means 2; and 3) one data input means. Voice input selection detecting means 3 for detecting voice input selection data, 4), voice input means 4, 5) voice features for extracting features of voice input by the voice input means 4 from input data Extraction means 5 and 6), voice features and data storage means 6 for storing a pair of data, 7) Input voice features by the voice feature extraction means 5 and voice features in the data storage means 6 As a voice feature pattern matching means 7 for performing those pattern matching
8) 1 data input means, 3 voice input selection detecting means, 6 data storage means, 7 voice feature pattern matching means as input, and 2 data output means as output And a data input / output control means 8 for controlling the input / output of the voice input device and the voice input device.

[0010]

1 is a schematic view of the concept of the present invention. Also,
2, 3 and 4 are diagrams showing an outline of the processing in the present invention. The present invention will be described in detail below with reference to these four figures.

Most of the computers currently in use have a keyboard independent of the main body. Alternatively,
In many cases, the terminal device is independent of the computer that performs the main processing. In these cases, each has a CPU independent of the main body, and executes processing by communicating with a computer that performs main processing.

The present invention provides a voice input device which is as easy to use as a widely used "kana-kanji conversion Japanese input" by connecting a voice recognition device using the hardware and software connected to the communication line. It will be realized. Of course, the connection to the communication line in the present invention does not necessarily have to be a physical connection. For example, it may be a software connection via a device driver of the computer main body.

The operation of the present invention will be described with reference to FIGS. 1 and 2. First, the voice input selection data is registered in advance.
It is conceivable that this selection data uses non-printing characters, for example, control character data. After that, the data inputted from the data input means 1 is sent to the voice input selection detecting means 3. If the input data is not the voice input selection data, the data is sent to the data output means 2 via the data input / output control means 8 and thereby to the computer main body which performs main processing. This process is exactly the same as the normal case where the keyboard and the terminal device are directly connected to the computer main body.

If the data sent to the voice input selection detecting means 3 is the voice input selection data,
The data itself is discarded, and as a side effect thereof, the voice input means 4 and the voice feature extraction means 5 are activated. Then, in this state, the features of the voice input from the voice input unit 4 are extracted by the voice feature extraction unit 5. Then, the process 1 shown in FIG. 3 is started.

In the process 1, the voice feature extracted by the voice feature extraction means 5 and the voice feature in the data storage means 6 are collated in the voice feature pattern matching means 7.

If the characteristic of the input voice can be found in the data storage means 6, the data registered in the data storage means 6 as a pair of the characteristic patterns is output as data 2. Output to the means. For example, it is a data string corresponding to a character string such as "Yamato 3-3-5, Suwa City, Nagano Prefecture" for the voice "Address". In this case, if a program such as a word processor is running on the computer that performs the main processing, this will give the same result as the above character string being input at the cursor position at the time the voice was input. give. In other words, it is possible to realize voice input without modifying the application software. The data paired with the voice feature does not necessarily have to be the data corresponding to the “character”. For example, it is naturally possible that the data is not "character" such as control data for realizing a certain function.

If the feature of the input voice cannot be found in the data storage means 6, the pre-registered data corresponding to the unknown pattern, for example,
The data corresponding to the characters such as "?????????" is also output to the data output means 2.

Normally, the output data string is echoed back on the CRT, and the user confirms it and decides whether or not to confirm the input. If the user performs the input confirmation processing, this voice input is determined and the line feed data or the like is output. If the user does not perform the confirmation process, the process 2 for correcting the character string shown in FIG. 4 is started.

Process 2 is essentially a normal character string modification process. For example, when "????????" is output for "address", the user uses a cursor key or the like as in a normal correction operation, and the character string is changed to "3 Yamato, Suwa-shi, Nagano Prefecture".
-3-5 ", etc., and confirm operation. Then, the system registers this corrected character string in correspondence with the voice pattern just input.

[0020]

As described above, according to the present invention, the voice input can be realized by adding the minimum voice input system to the existing hardware and without making any change to the software. it can. Further, the above operation is equivalent to the widely used "kana-kanji conversion Japanese input", which is easily accepted by the user. Also in the voice recognition processing, according to the method of the present invention, the user can easily switch between the voice input and the keyboard input, no erroneous segmentation of the utterance section occurs, and the load of computer resources is extremely high. It is a small one.

In the present invention, the hardware- and software-added portion of the present invention need not be independent of the computer main body. This can also be realized as a software keyboard device driver program using the hardware included in the computer body.

[Brief description of drawings]

FIG. 1 is a schematic diagram of the concept of the configuration of the present invention.

FIG. 2 is a flowchart showing a schematic algorithm of the whole processing of the present invention.

FIG. 3 is a flowchart showing a schematic algorithm of a voice feature pattern matching process of the present invention.

FIG. 4 is a flowchart showing a schematic algorithm for data correction of the present invention.

[Explanation of symbols]

 1: Data input means 2: Data output means 3: Voice input selection detection means 4: Voice input means 5: Voice feature extraction means 6: Data storage means 7: Voice feature pattern matching means 8: Data input / output control means

Claims (1)

[Claims] 1. A voice input selection data is detected from data inputted by 1) one or more sets of data input means, 2) one or more sets of data output means, and 3) one data input means. 4), voice input means, 5) voice feature extraction means for extracting features of voice input by the voice input means, 6), voice feature and data pair 7) a data storage means for storing the following: 7) a voice feature pattern matching means for performing a pattern matching of the voice features by the voice feature extraction means of 5 and the voice features in the data storage means of 6 as inputs; ) 1 data input means, 3 voice input selection detection means, 6 data storage means, 7 voice feature pattern matching means as input, and 2 data output means as output, input / output of data A voice input device comprising: a data input / output control means for controlling force;