JP6712940B2 - Voice input device, voice input method - Google Patents

Description

The present invention relates to a voice input device and a voice input method.

Application data is created by inputting character strings or the like to input items on application screens for various services from a touch panel of a smartphone or an input device such as a keyboard connected to a PC (personal computer). .. Some application screens have a plurality of input items, for example, input items such as name, address, and telephone number.
When inputting data for such an input item, the user touches the input field of the input item to be input on the screen or operates the mouse to move the pointer to the input field and click. , Select which input item to enter. Then, the data is input from the touch panel or the keyboard.
As a method of inputting data for such an input item, there is a method of inputting a voice by using a method other than a method of inputting a character string from an input device (for example, refer to Patent Document 1).

JP, 2014-085954, A

However, when there are a plurality of input items, the user speaks after specifying the input item to be input by touch input, etc., and then touches the next input item to specify and speak. It is necessary to repeat the operation, and as the number of input items increases, the number of operations for designating an input item to be input by performing a touch operation or the like increases and the time and effort required for input increases.

The present invention has been made in view of such circumstances, and an object thereof is a voice input device capable of preventing an increase in time and effort required for operation input when voice input is performed for a plurality of input items. , Providing a voice input method.

In order to solve the above-mentioned problems, the present invention detects a plurality of keywords included in a voice recognition unit that generates text data from a voice signal corresponding to a collected voice, and the keywords included in the text data, and detects the detected keywords. Among them, the text data between the first keyword and the second keyword is extracted as the fragment text data for the first keyword, and at least a part of the text data in the latter part of the second keyword is extracted as the fragment text data for the second keyword. An input data extraction unit for extracting, a word that is a candidate for the divided text data for the input item, and a dictionary data storage unit that stores at least one of a plurality of small items that make up the input item in association with each other, and an input operation. The input data extraction unit obtains the text data to be input to the input field for the input field on the display screen in which the input target data including a plurality of input fields in which the text data can be input is displayed. was acquired from split the text data has a corresponding data input unit for inputting to each of the input field, wherein the input data extraction unit, the dictionary data storage unit to the detected keyword from dividing keyword If there is a stored word, specify the character string immediately after the keyword as the target character string to be divided into small items even if the text data does not include the name of the small item. , The data input unit has a character string to be divided into the small items in an input field having a correspondence relationship with the name of the small item corresponding to the word among the names of the small items displayed on the display screen. Describe the word obtained based on the string specified to be .

Further, the present invention is, in the above-described voice input device, a conversion data storage unit that stores conversion data corresponding to the input item, and the conversion data in at least a part of the divided text data extracted by the input data extraction unit. A data conversion unit that converts the conversion target character into a character different from the character based on the conversion data when the conversion target character stored in the storage unit is included, The input unit describes, for an input item, the divided text data after the divided text data extracted by the input data extraction unit has been converted by the data conversion unit.

Further, the present invention is a voice input method in a computer, wherein the voice recognition unit generates text data from a voice signal corresponding to the collected voice, and the input data extraction unit is a keyword included in the text data. Of the detected keywords, the text data between the first keyword and the second keyword is extracted as the divided text data for the first keyword, and at least a part of the text data of the latter stage of the second keyword is detected. From the divided keywords to the second keyword, and the input data extracting unit extracts, in the keyword detected from the divided keywords, a word that is a candidate for the divided text data for the input item and a plurality of small words that form the input item. When there is a word stored in the dictionary data storage unit that stores at least one of the items in association with each other, a character string immediately behind the relevant keyword is used, and the text data includes the name of the small item. Even if it is not specified, it is specified as the target character string to be divided into small items, and the data input part displays the input target data that includes multiple input fields that can input text data according to the input operation. the input field, text data for input to the input field, is obtained from divide text data obtained by the input data extraction unit, and input to the corresponding said input field, the data input unit Is specified as a character string to be divided into the sub-items in the input field corresponding to the name of the sub-item corresponding to the word among the names of the sub-items displayed on the display screen. Describe the word obtained based on the character string .

As described above, according to the present invention, voice input can be performed without increasing the time and effort required for operation input when voice input is performed for a plurality of input items.

It is a schematic block diagram which shows the structure of the portable terminal which applied the audio|voice input device by one Embodiment of this invention. It is a figure which shows an example of the input screen of the input object in the portable terminal 1. 6 is a flowchart illustrating an operation of the mobile terminal 1. It is a conceptual diagram explaining the process which acquires the input data with respect to an input item. It is a figure showing an example of a screen showing a state after a character string is inputted to an input item.

Hereinafter, a voice input device according to an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a schematic block diagram showing the configuration of a mobile terminal to which a voice input device according to an embodiment of the present invention is applied. The mobile terminal 1 includes a voice signal generation unit 11, a voice recognition unit 12, a dictionary data storage unit 13, a conversion data storage unit 14, an input data extraction unit 15, a data conversion unit 16, a data input unit 17, a display unit 18, and an operation input. It has a part 19.

The voice signal generator 11 collects voice and generates a voice signal. For example, a microphone is used as the audio signal generation unit 11. The voice recognition unit 12 generates text data from the voice signal generated by the voice signal generation unit 11.

The dictionary data storage unit 13 stores a word that is a candidate for input data for an input item and at least one of a plurality of small items that make up the input item in association with each other. Here, the input data is at least a part of the text data, and is data that can be input to the input field of the input item. Moreover, the input item may include a plurality of small items. For example, when the input item is "address", there are "prefecture", "city", "street number", "building name" (or "street number" etc.), etc. as small items. One input item is composed of the small items. As words that can be candidates for input data, for example, in the case of “prefecture”, names that can exist as prefectures such as “Tokyo”, “Osaka prefecture”, “Hokkaido”, “Saitama prefecture” are stored. It

The conversion data storage unit 14 stores conversion data corresponding to the input item. Here, although the characters represented by the text data can be directly applied as the input data, the general utterance content and the general description content may be different in the input item. For example, when speaking between the area code and area code of the telephone number, and between the area code and the subscriber number, speak as "no" such as "03, 0000, 0000", but the telephone number is converted to a character string. In describing, "-" (hyphen) is used like "03-0000-0000". In this way, even if the utterance content is different from the written character string, even if the user utters without being aware of the written character string, it can be converted into a character string according to the writing method. It will be possible.

The input data extraction unit 15 detects a plurality of keywords included in the text data, and among the detected keywords, the text data between the first keyword and the second keyword is used as the input data (divided text data) for the first keyword. It is extracted and extracted as input data (divided text data) for the second keyword from at least a part of the text data in the latter part of the second keyword. As described above, based on the keyword, it is possible to distinguish which of the plurality of input items is the character string to be input by delimiting it. Details will be described later.
When at least a part of the input data extracted by the input data extraction unit includes the conversion target character stored in the conversion data storage unit, the data conversion unit 16 determines the conversion target character based on the conversion data. Converts a character to a different character.

The data input unit 17 sets the text data to be input in the input field to the input field in the display screen on which the input target data including a plurality of input fields in which the text data can be input according to the input operation is displayed. , The text data generated by the voice recognition unit is obtained from the divided text data obtained by dividing the text data, and is input into the corresponding input fields. In addition, when the input data extracted by the input data extraction unit 15 has a word stored in the dictionary data storage unit 14, the data input unit 17 describes the word in a small item corresponding to the word. Further, the data input unit 17 describes the text data after the input data extracted by the input data extraction unit 15 has been converted by the data conversion unit 16 in the input item.

The display unit 18 is, for example, a liquid crystal panel or the like, and displays various information. The operation input unit 19 is a touch sensor and detects which position in the detection area is touched. A touch panel can be used as the display unit 18 and the operation input unit 19. The control unit 20 controls each unit in the terminal device 1.

FIG. 2 is a diagram showing an example of an input screen of an input target on the mobile terminal 1.
The input screen may be, for example, an input screen that includes a plurality of pairs of input items and input fields corresponding to the input items. For example, various services (bank account opening, loan, travel, credit card creation, etc.) There are application data, creation request data for requesting the creation of a quotation, and destination data for registering the delivery destinations of products in online shopping. On this input screen, there are "name" (reference numeral 100), "address" (reference numeral 110), "phone number" (reference numeral 120), and "company name" (reference numeral 130) as input items. The input item "address" further includes a plurality of input items as small items. Here, "prefecture" (reference numeral 111), "municipalities" (reference numeral 112), "address etc." (reference numeral 113). There is a small item. An input field (reference numeral 140) is arranged in each of these input items or small items. Further, a voice input button (reference numeral 200) for switching on and off of a function for performing voice input is provided on a part of the input screen (here, the upper right of the screen).

Next, the operation of the above-described mobile terminal 1 will be described with reference to FIGS. 3, 4, and 5. 3 is a flow chart for explaining the operation of the mobile terminal 1, FIG. 4 is a conceptual diagram for explaining a process for obtaining input data for an input item, and FIG. 5 is a state after a character string is input for the input item. It is a figure which shows the example of a screen showing.
First, the control unit 20 of the mobile terminal 1 displays the application data to be input (for example, FIG. 2) on the screen of the display unit 18 based on the operation input from the user. The control unit 20 detects whether or not a touch input has been made on the screen displayed on the display unit 18 (step S101). If no touch input has been made (step S101-NO), wait for a certain period of time (step S102), and the process proceeds to step S101.

On the other hand, when the touch input is performed, the control unit 20 determines whether or not the instruction is to turn on the voice input (step S103). Whether or not the instruction is to turn on the voice input can be determined based on whether or not the operation input is performed by touching the voice input button (reference numeral 200 in FIG. 2). If it is not an instruction to turn on voice input (step S103-NO), for example, if a character is input by touching an input target item and then touching an operation button on the touch panel, a character string by touch input Is input (step S104).

On the other hand, when the operation is to turn on the voice input, the control unit 20 starts the voice input process. When the voice input process is started, the voice signal generation unit 11 acquires the voice uttered by the user and generates a voice signal corresponding to the obtained voice (step S105). When the voice signal is generated, the voice recognition unit 12 performs voice recognition processing based on the voice signal to generate text data corresponding to the voice signal (step S106). For example, the text data obtained here is “Name is Taro Yamada, Address is 1 in 1 of Minato-ku, Tokyo, Phone number is 03, 0000, 0000 Company name is Patent Corporation” (FIG. 4, symbol 300). ..

When the text data is generated, the input data extraction unit 15 detects a keyword included in the text data (step S107). The keyword detected here is a character string corresponding to the name of the input item. For example, "name" (reference numeral 400 in FIG. 4), "address" (reference numeral 402 in FIG. 4), and "phone number" (reference numeral 404 in FIG. 4). ), and “company name” (reference numeral 406 in FIG. 4). Then, the input data extraction unit 15 specifies the character string between the keywords based on the detected plurality of keywords and also specifies the character string of the latter stage of the final keyword (step S108). Here, the final keyword is the keyword at the end of the sentence of the obtained text data, and is the “company name” here. Then, in this step S108, the character string specified as the character string between the keywords is, for example, “Taro Yamada”, “is 1 to 1 of Minato-ku, Tokyo”, or “is 03 to 0000”. The character string specified as the character string following the final keyword is “wa patent corporation”. If a specified character (for example, "ha") is at the beginning of the specified character string, it is excluded from the specified character string, and "Yamada Taro" (reference numeral 401 in FIG. 4), "Minato Ward, Tokyo" The character strings "1 in 1 of 1" (reference numeral 403 in FIG. 4), "0000 of 0000 in 03" (reference numeral 405 in FIG. 4), and "patent corporation" (reference numeral 407 in FIG. 4) are obtained as the specified character strings.

Next, the input data extraction unit 15 divides the character string for an input item having a small item so as to correspond to the small item (step S109). Regarding whether or not the small item is an input item, the input data extraction unit 15 registers the input item name in a predetermined memory area of the mobile terminal 1 in advance, and stores the input item name in the memory detected by the keyword detected in step S107. It is determined whether there is a keyword corresponding to the input item name registered in the area, and if there is a corresponding keyword, the character string immediately behind that keyword is the character string to be divided into sub-items. Identify as. Then, the input data extraction unit 15 refers to the dictionary data storage unit 13 and determines whether or not the identified character string includes the character string registered in the dictionary data storage unit 13. For example, the dictionary data storage unit 13 is associated with the input item “address”, and the prefectures (for example, 47 prefectures such as “Tokyo”, “Kanagawa prefecture”, “Osaka prefecture”, etc.) are candidates for the address. Name), the name of the municipality (for example, the name of a municipality that can exist in 47 prefectures such as "Minato Ward", "Chiyoda Ward", "Aoba Ward", "Chuo Ward") The data is stored. The input data extraction unit 15 extracts the character strings “Tokyo” and “Minato Ward” stored in the dictionary data storage unit 13 from “1 of 1 in 1 Minato-ku, Tokyo” which is the character string after “Address”. And “1 of 1 in 1” are also specified as the character strings corresponding to the small items, and the original character string “1 in Minato-ku, Tokyo” is also specified. "1 of 1" is divided to obtain three character strings "Tokyo", "Minato Ward", and "1 of 1 of 1".

Next, the input data extraction unit 15 refers to the conversion data storage unit 14 and converts the character to be converted into another character for each of the specified character strings (step S110). Here, the conversion data storage unit 14 targets the character string corresponding to the small item “address, etc.” of the input item “address”, and when the character “no” is present, the converted character is “−”. (Hyphen) is memorized, and if there is a character "no" in the character string corresponding to the input item "phone number", the converted character must be "-" (hyphen). I remember. The input data extraction unit 15 refers to the conversion data storage unit 14, and if the character string has a character to be converted, converts the character. For example, the character string of "1 of 1" is converted into "1-1-1", and the character string of "03 of 0000-0000" is converted into "03-0000-0000".

Next, the data input unit 17 describes the character string obtained by the input data extraction unit 15 in the input field of the corresponding input item based on the keyword (step S111).
Here, for example, the data input unit 17 identifies the character string “Taro Yamada” between the keyword “name” and the keyword “address” as the input data for the input item “name”, and sets it as “name”. Describe in the corresponding input field. Similarly, the data input unit 17 divides the character string between the keyword “address” and the keyword “phone number” in step S109, and inputs the divided character string into the corresponding sub-item. Describe in the column. Here, the character string “Tokyo” is described in the small item “prefecture”, and “Minato Ward” is described in the small item “city/town/village”. Regarding the “address, etc.”, the character string after being converted in step S110. “1-1-1” is described. Further, the data input unit 17 inputs the character string “03-0000-0000” which is a character string between the keyword “telephone number” and the keyword “company name”, which is converted in step S110. It is specified that the input data is for "telephone number", and is described in the input field corresponding to "telephone number". Then, the data input unit 17 identifies the character string "patent corporation" in the latter part of the final keyword "company name" as the input data for the input item "company name", and inputs the corresponding "company name". Describe in the column. When the character string is described in the input field of the input item in this manner, as shown in FIG. 5, the character string based on the information input by voice is described in each input field. When the send button (reference numeral 500) is touch-inputted after this description is made, the mobile terminal 1 sends the input data for the input item to the server device to which the application data is sent.

In the embodiment described above, if there is a character string to be corrected in the described character string on the screen after the voice input shown in FIG. 5, touch the input field of the correction target. Then, the character string can be corrected by inputting a voice or a character string by a touch operation to the input field.

Further, in the above-described embodiment, the case where the voice input process is started from the time when the voice input button is touched in step S103 has been described. However, as the end timing of the voice input process, the voice input button is touched again. Until a predetermined time elapses after the voice input button is touched, or until the touched finger is separated from the voice input button.

Further, in the above-described embodiment, the case where one voice input button is provided on the application data input screen has been described, but a plurality of voice input buttons may be provided as long as the number is smaller than the number of input items. .. For example, a voice input button for voice inputting a name and an address and a voice input button for voice inputting a telephone number and a company name may be provided.

Further, in the above-described embodiment, in the process of identifying a character string in step S108, if a specified character (for example, “ha”) is present at the beginning of the identified character string, it is excluded from the identified character string. However, when "ga" is registered in advance as a character to be excluded and a character string "ga Yamada Taro" is obtained from text data such as "Name is Taro Yamada...", “Ga” may be excluded as the predetermined character at the beginning of the character string.

Moreover, although the case where the voice input process according to the present invention is applied to the mobile terminal 1 has been described in the above-described embodiment, it may be applied to a smartphone or a PC (personal computer). Also, voice input is performed in the terminal device, the obtained voice data or text data is transmitted to the server device, the voice input processing according to the present invention is performed in the server device, and the result after being described in the input field is displayed in the terminal. It may be transmitted to the device. Further, for example, among the functions of the mobile terminal 1, the functions of the voice recognition unit 12, the dictionary data storage unit 13, and the conversion data storage unit 14 may be performed in the server device.

The functions of the voice signal generation unit 11, the voice recognition unit 12, the dictionary data storage unit 13, the conversion data storage unit 14, the input data extraction unit 15, the data conversion unit 16, and the data input unit 17 in the above-described embodiment are realized by a computer. You may do it. In that case, the program for realizing this function may be recorded in a computer-readable recording medium, and the program recorded in this recording medium may be read by a computer system and executed. The “computer system” mentioned here includes an OS and hardware such as peripheral devices. Further, the "computer-readable recording medium" means a portable medium such as a flexible disk, a magneto-optical disk, a ROM, a CD-ROM, or a storage device such as a hard disk built in a computer system. Further, the "computer-readable recording medium" means to hold a program dynamically for a short time like a communication line when transmitting the program through a network such as the Internet or a communication line such as a telephone line. In this case, a volatile memory inside a computer system that serves as a server or a client in that case, which holds a program for a certain period of time, may be included. Further, the program may be for realizing a part of the functions described above, or may be a program that can realize the functions described above in combination with a program already recorded in a computer system, It may be realized using a programmable logic device such as FPGA (Field Programmable Gate Array).

Although the embodiment of the present invention has been described in detail above with reference to the drawings, the specific configuration is not limited to this embodiment, and includes a design and the like within a range not departing from the gist of the present invention.

DESCRIPTION OF SYMBOLS 1... Mobile terminal 11... Voice signal generation part 12... Voice recognition part 13... Dictionary data storage part 14... Conversion data storage part 15... Input data extraction part 16... Data conversion part 17... Data input part 18... Display part 19... Operation Input section

Claims (3)

A voice recognition unit that generates text data from a voice signal corresponding to the collected voice,
A plurality of keywords included in the text data is detected, and among the detected keywords, text data between a first keyword and a second keyword is extracted as divided text data for the first keyword. An input data extraction unit that extracts, as at least a part of the text data of, as divided text data for the second keyword,
A dictionary data storage unit that stores a word that is a candidate for segmented text data for an input item and at least one of a plurality of small items that form the input item in association with each other;
With respect to the input field on the display screen on which the input target data including a plurality of input fields in which the text data can be input according to the input operation is displayed, the text data to be input in the input field is input to the input data extraction unit acquired from divide text data obtained by having a data input unit for inputting to the corresponding said input field, a,
When the keyword detected from the divided keywords has a word stored in the dictionary data storage unit, the input data extraction unit assigns a character string immediately behind the keyword to the name of a small item in the text data. Even if it does not include, specify that it is a character string to be divided into small items,
The data input unit, in the input field having a correspondence relationship with the name of the sub-item corresponding to the word among the names of the sub-items displayed on the display screen, the character string to be divided into the sub-items. A voice input device that describes a word obtained based on a specified character string . A conversion data storage unit that stores conversion data corresponding to the input item;
If at least a part of the divided text data extracted by the input data extraction unit includes a conversion target character stored in the conversion data storage unit, the conversion target character is based on the conversion data. And a data conversion unit that converts the character into a character different from the character,
The voice input device according to claim 1, wherein the data input unit describes, for an input item, the divided text data after the divided text data extracted by the input data extraction unit has been converted by the data conversion unit. A voice input method in a computer,
The voice recognition unit generates text data from a voice signal corresponding to the collected voice,
The input data extraction unit detects a plurality of keywords included in the text data, and of the detected keywords, extracts text data between a first keyword and a second keyword as divided text data for the first keyword, Extracted as segmented text data for the second keyword from at least a part of the text data in the latter stage of the second keyword,
The input data extraction unit stores the keyword detected from the split keyword in association with a word that is a candidate for the split text data for the input item and at least one of a plurality of small items that form the input item. When there is a word stored in the dictionary data storage unit, the character string immediately after the keyword is to be divided into sub-items even if the text data does not include the name of the sub-item. Is identified as
The data input unit, for the input field in the display screen on which the input target data including a plurality of input fields in which the text data can be input according to the input operation is displayed, the text data to be input in the input field, acquired from divide text data obtained by the input data extraction unit, and input to the corresponding said input field,
The data input unit, in the input field having a correspondence relationship with the name of the sub-item corresponding to the word among the names of the sub-items displayed on the display screen, the character string to be divided into the sub-items. A voice input method that describes a word obtained based on a specified character string .

Images