JP2009163528A - Commodity sales data processing apparatus, program thereof, and commodity data input apparatus and program thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a commodity sales data processing apparatus that extracts commodity categories or commodity character strings similar to a recognized voice as sound without omission, even when the recognized voice is erroneously recognized, and that outputs and displays commodity data set according to commodity category candidates and commodity character string candidates without omission on the basis of the candidates. <P>SOLUTION: The commodity sales data processing apparatus includes: a commodity category/commodity data dictionary 1131 and a commodity character string/commodity data dictionary 1132 which store commodity categories and commodity character strings set correspondingly to commodity data for identifying respective commodities respectively; a voice pattern data extraction means 111 for extracting a commodity category or a commodity name character string which is coincident with or similar to voice data uttered by an operator as sound; a commodity category candidate storage means 1141 and a commodity character string candidate storage means 1142 which store these extracted candidates respectively; and an commodity data reading means 1113 for reading commodity data set correspondingly to these stored candidates. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a merchandise sales data processing apparatus using voice recognition technology, a program that causes the apparatus to function, a merchandise data input apparatus, and a program that causes the apparatus to function.

In recent years, data input using voice recognition technology that uses voice instead of a keyboard has been put into practical use in various situations, and product sales data processing devices, product data input devices, etc. using this technology are known. Yes. These devices include keyword storage means for presetting and storing a keyword for specifying a group to which the product belongs for each product. When an operator inputs a specific keyword by voice, the keyword is stored based on the keyword. Correspondingly set product names are extracted and displayed on the screen (for example, see Patent Document 1).
JP 2000-293754 A

  However, with current voice recognition technology, there is a risk of misrecognition unless the correct name is spoken, and there is a possibility that recognition may fail due to the influence of ambient noise and other people's conversation. For this reason, when the product name is long or a similar product name is generated, the user burden for correctly recognizing the voice is large. In addition, since the above-mentioned device is not provided with a case where the input speech is misrecognized, the product name belonging to the wrong group may be output and displayed on the screen when the wrong keyword is recognized. It is done. For this reason, there is a possibility that a candidate omission such as the display of the product name desired by the operator may not occur, and there is a possibility that the sales data processing takes time or the processing cannot be performed.

  The present invention has been made based on such circumstances, and misrecognized speech is considered to be at least speech-similar. Therefore, speech-similar candidates are extracted without omission and An object of the present invention is to provide a merchandise sales data processing apparatus capable of outputting and displaying voice candidates desired by an operator even when there is a misrecognition of input voice by outputting and displaying corresponding merchandise data.

  Corresponding to an acoustic dictionary that stores voice feature values and voice pattern data that are created in advance in association with each other, a voice dictionary that stores the voice pattern data in association with product categories, product data that identifies products, and this product data A product category / product data dictionary storing the product category set as described above, a voice input unit for inputting voice, and a voice feature amount of voice input by the voice input unit with reference to the acoustic dictionary; A speech recognition unit that compares speech feature values created in advance and outputs the speech pattern data stored in association with a matching or similar speech feature amount; and by referring to the speech dictionary, the speech recognition unit Voice pattern data extracting means for extracting the product category as a candidate based on the output voice pattern data; Product category candidate storage means for storing the product category extracted by the voice pattern data extraction means as product category candidates, and the product data set corresponding to the product category candidate from the product category / product data dictionary Product data reading means to be read, output display means for outputting and displaying the product data read by the product data reading means, and arbitrary product data from the product data output and displayed by the output display means A product sales data processing apparatus comprising: product sales data processing means for performing product sales data processing when selected.

  According to the present invention, when a part of a character string of a product category or product name is input by voice, voice candidates such as a plurality of product category candidates and product character string candidates determined to be similar to the input voice are extracted. Since these voice candidates are set corresponding to each piece of product data, the product data corresponding to the product category candidates and product character string candidates can be output and displayed without omission. Thereby, it is possible to prevent candidate omission when the input voice is erroneously recognized. Further, the sales process or the input process can be performed by speaking a part of the product category or the product character string. For this reason, there is an effect that it is possible to operate the apparatus even when the operator does not know the official product name at the time of the sales input processing operation.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In addition, this embodiment is a case where it applies to the processing apparatus used for the merchandise sale for restaurants.

  In the first embodiment, when an operator inputs a product category by voice, a product category candidate that matches or resembles the voice pattern data of the product category is referred to and set corresponding to the product category candidate. A merchandise sales data processing apparatus that outputs and displays merchandise data, and will be described with reference to FIGS. The product category in the present embodiment refers to an attribute that specifies a group to which each product belongs for each product. FIG. 1 is a perspective view showing an appearance of a product sales data processing apparatus according to the first embodiment of the present invention. The merchandise sales data processing apparatus 1 is provided, and a microphone 13 is provided as a voice input means for inputting an operator's voice. Something is used. However, the microphone 13 does not need to be provided separately from the main body as illustrated, and may be built in the output display means 12. As for the output display means 12, a touch panel system is used in the present embodiment, but it is not necessary to be limited to this.

  Next, FIG. 2 shows an internal configuration of the apparatus main body 11. In the apparatus main body 11, a CPU (Central Processing Unit) 111 that controls each apparatus and performs the core function of the computer, and a ROM (Read Only Memory) 112 in which fixed data such as a program for controlling the operation of the CPU 111 is stored in advance. , RAM (Random Access Memory) 113 in which data is written and read directly from the CPU 111, HDD (Hard Disc Drive) 114, which is a large-capacity storage medium capable of storing sales data and various data, LAN (Local A communication interface 115 that controls data communication with an electronic device connected via a communication network such as an area network), a display controller 116 that controls screen display on the display, and a touch panel interface that receives touch detection signals from the touch panel sensor 122. 117, input from microphone 13 Speech recognition engine 119 to perform speech recognition based on the digital audio data outputted from the A / D converter 118 and the A / D converter 118 converts analog audio data to digital audio data is provided with. The CPU 111 and ROM 112, RAM 113, HDD 114, communication interface 115, display controller 116, touch panel interface 117, and speech recognition engine 119 are electrically connected by a bus line such as an address bus or a data bus.

  Subsequently, the voice input means captures the voice from the start to the end of the operator's voice input via the microphone 13, and corresponds to steps ST9-1 to ST9-4 in FIG. . When the voice capture method detects that the voice recognition key 37 in FIG. 3 is touched by a signal from the touch panel sensor 122, the CPU 111 converts the analog voice data into digital voice data by A / D conversion. The voice recognition engine 119 takes in the voice data. Note that the capturing of the voice data is continued until the finger is released from the voice recognition key 37 on the display 121 and turned off. Then, noise processing is performed after A / D conversion.

  A noise processing method of the voice input means will be described. Noise processing includes a delay sum method using a directional microphone and an HMM synthesis method in which an acoustic dictionary 1134 in which noise is previously included in speech pattern data is prepared to perform noise processing. In the present invention, the spectrum subtraction method is used. Noise processing by (hereinafter referred to as SS method) is used.

  The spectrum subtraction method (hereinafter referred to as SS method) is a method of subtracting the noise amplitude spectrum from the amplitude spectrum of the voice data produced by mixing the noise into the voice uttered by the operator or from the power spectrum of the voice data containing the noise. Noise suppression is achieved by subtracting the power spectrum. The power spectrum is a square of the amplitude spectrum, and the output by the SS method is an amplitude spectrum or a power spectrum in which noise is suppressed. However, the noise processing in the implementation of the present invention is not limited to the SS method, and any method may be used as long as it can process the noise and extract voice data.

  Next, the voice recognition unit refers to the acoustic dictionary 1134 based on the voice feature amount of the voice captured during the period when the voice recognition key 37 is turned on by the CPU 111 operating the voice recognition engine 119. It is a means for outputting voice pattern data that matches or is similar. Note that the voice pattern data here indicates a phoneme pattern string output from an acoustic dictionary 1134 described later. Specifically, the speech feature amount is obtained by performing linear prediction analysis based on the input sound. Linear prediction analysis is a technique for obtaining the spectral envelope from the input sound, and is a commonly known speech feature extraction method that reflects the vocal tract characteristics of the utterance mechanism (Kiyohiro Shikano (4 others) "Speech recognition system" ohm Publication, first edition (May 2001) P1-P13). The speech recognition means outputs speech pattern data using an acoustic dictionary 1134 that stores speech feature quantities and speech patterns created in advance. For example, the HMM method (referred to by Seiichi Nakagawa, “Speech recognition using a probabilistic model”, Electronic Information and Communication Engineers P29 to P80) is used for outputting voice pattern data.

  The voice pattern data extraction unit 1111 is a unit that extracts a product category as a product category candidate based on the voice pattern data output by the voice recognition unit. Candidates for reading product data are extracted by this means and stored in the product category candidate storage means 1141 described later.

  The CPU 111 reads the product data set corresponding to the product category candidate from the product category / product data dictionary 1132 described later, the product data input process that has been sold, or the registration process. Product sales data processing means 1114 for performing sales processing is provided.

  The product data read by the product data reading unit 1113 is displayed on the display 121. Here, an example of the product sales processing screen 38 displayed on the display 121 is shown in FIG. As shown in the figure, the product sales processing screen 38 displays touch keys 35 such as a confirmation key and a cancel key in addition to the numeric keys 33 such as “0” to “9”. Further, at the upper left of the product sales processing screen 38, there is a table 31 for displaying products, quantity, and price, and a table 32 for displaying inputted numerical values. Here, the operator speaks while pressing the voice recognition key 37 on the touch panel, and takes in the voice. After the audio capture is completed, the product data obtained from the recognized speech candidates is displayed in the table 36 located in the center of the product sales processing screen 38. The product data item selected by the operator is displayed on the display unit 34 located at the upper center of the screen.

  Next, an example in which product category candidates are output on the product sales processing screen 38 will be described with reference to FIG. In the present embodiment, about 10 pieces of product data obtained from voice candidates are output on the display 121 at the maximum. However, it is not necessarily limited to this form, and all the product data set corresponding to the extracted product category candidates may be output on the product sales processing screen 38, or the product data is sequentially ordered in the scroll format. May be displayed. Further, as the display format, only the name of the product may be shown, or the product image may be displayed simultaneously with the name. Further, additional information such as merchandise sales information may be added to display more detailed merchandise information, and the display format is not limited to this embodiment.

  Next, when displaying the product data read by the product data reading unit 1113 on the display 121, the output display control unit 1161 performs sorting processing based on the information of the products. . For example, when the product data information is the number of times a product has been selected and input, the items are rearranged in the order of products that have been frequently input. Further, when the product data information is the product price, the ordering control may be performed with a certain regularity such as rearrangement in the order of price, so that the operator can easily perform the processing operation.

  A dictionary stored in the RAM 113 will be described. The RAM 113 includes a product category / product data dictionary 1131, a product data / product character string dictionary 1132, an audio dictionary 1133, and an acoustic dictionary 1134. The acoustic dictionary 1134 shown in FIG. 5 is associated with a voice feature amount of voice and voice pattern data. For example, when the word spoken by the operator is “Anpan”, the sound feature vector of this sound is shown as 51 in FIG. 5, and the sound pattern data 52 is corresponding to “anpan”. It is stored in association. Next, the speech dictionary 1133 shown in FIG. 6 stores speech pattern data 52 describing speech reading when a predetermined word is uttered. This voice pattern data 52 is a way of reading for each product category here. The voice pattern data 52 and the product category 61 are stored in association with each other, and what product category 61 is recognized by the voice dictionary 1133 is recognized.

  The product category / product data dictionary 1131 is a means for storing product data for identifying products and product categories set corresponding to the product data. FIG. 7 shows data contents stored in the product category / product data dictionary 1131. In the figure, a product category 61, a menu code 72, and a product name 73 as product data are stored as data, and the same product category 61 is set for the product names 73 classified into the same product category 61. . Based on the product category 61, the product data reading unit 1113 reads the product name 73 set as the same product category 61. Here, the product data may be not only the product name 73 but also image information or sales information.

  For example, “Anpan” is set as the product category 61 corresponding to the product name 73 “Muanpan”. At this time, the image information of “Muanpan” together with the product name 73 may be set as product data in association with the product category 61 of “Anpan”. The product category may be classified according to appearance features such as the color, shape, and pattern of the product. For example, in the case of bread, in addition to general product name categories such as “Anpan” and “bread”, color categories such as “black” and “white”, “circle”, “square”, “ It may be a shape category such as “triangle” or a pattern category such as “mottle” or “lattice”. These merchandise categories 61 are not limited to the classifications exemplified above, and may be anything that can classify merchandise. The product data / product character string dictionary 1132 will be described in a second embodiment to be described later.

  In addition to the RAM 113, the above-described dictionary may be provided in the HDD 114 in the merchandise sales data processing apparatus, and the voice pattern data may be read from the HDD 114, or a dictionary may be provided in the server via an electric communication line. Data may be read from the server.

  Next, the HDD 114 includes a product category candidate storage unit 1141, a product character string candidate storage unit 1142, and a product sales data processing frequency information storage unit 1143. The merchandise category candidate storage means 1141 is a means for memorizing the voice category data candidate candidates extracted by the voice pattern data extracting means 1111. FIG. 8 shows the data contents stored in the product category candidate storage unit 1141, and the product category candidates extracted by the voice pattern data extraction unit 1111 are stored in the product category storage candidate area 81. The product sales data processing count information storage means 1143 is means for storing product sales data processing count information. The product sales data processing count may be, for example, a simple product selection input count, a sales registration count, or data regarding the product sales processing count. These storage means may be provided on the telecommunication line in the same manner as other dictionary / storage means.

  First, the merchandise sales data processing apparatus is activated when the power is turned on, and the CPU 111 displays a merchandise sales processing screen 38 that uses voice recognition during the merchandise sales processing shown in FIG. When moving to this screen, voice input is started, and ambient noise is started to be taken in from the microphone 13 attached to the apparatus main body 1 (ST9-1). When the operator speaks while pressing the voice recognition key 37 on the merchandise sales processing screen 38, the voice of the operator is captured (ST9-2). The captured voice is captured via the microphone 13 and A / D converted from analog voice data to digital voice data by the A / D converter 118 (ST9-3). The digital audio data after A / D conversion is subjected to noise processing by the SS method (ST9-4), and the noise contained in the digital audio data is removed.

  In this embodiment, the input section detection unit 1191 that detects the start and end of voice input is used to increase the accuracy of the recognition rate of voice recognition when the operator's voice is captured. This means is a means in which voice acquisition is started when the operator starts speaking and simultaneously presses the voice recognition key 37 displayed on the product sales processing screen 38. When the touch panel sensor 122 detects that the voice recognition key 37 in FIG. 3 is touched, the voice input by the operator is converted into digital voice data by the A / D converter 118, and the CPU 111 converts the digital voice data. The voice recognition engine 119 takes it in (ST9-2). The capturing of the digital voice data is continued until the finger is released from the voice recognition key 37 and turned off. Then, when the CPU 111 detects an operation of turning off the voice recognition key 37, the voice capturing ends.

After the completion of the voice acquisition, the voice features extracted during the period when the voice recognition key 37 is turned on by the voice recognition engine 119 are extracted (ST9-5), and the voice of the voice input with reference to the acoustic dictionary 1134 is extracted. The feature quantity is compared with the voice feature quantity created in advance, and it is determined whether or not they match (ST9-6). If it does not match or resembles the previously created speech feature (NO in ST9-6), the speech recognition is terminated (ST9-17), and if it matches or resembles the speech feature (YES in ST9-6), Based on the voice feature quantity, the acoustic dictionary 1134 is referred to, and voice pattern data stored in association with the matched or similar voice feature quantity is output (ST9-7).
Next, the voice dictionary 1133 is referred to based on the output voice pattern data, and it is determined whether there is a product category stored in association with the voice pattern data (ST9-8). For example, if the voice pattern data output by the voice recognition means at this time are “anpan” and “anman”, the voice pattern data is compared with the voice pattern data stored in the voice dictionary 1133 (ST9). -8). If the output voice pattern data does not exist in the voice pattern data in the voice dictionary 1133 with reference to the voice dictionary 1133 (NO in ST9-8), the voice recognition ends (ST9-17). Alternatively, the operator may be warned such as displaying an error message with a content “Please input by voice in another product category” or generating a warning sound. On the other hand, when the voice pattern data output by the voice recognition unit in ST9-8 exists in the voice dictionary 1133 (YES in ST9-8), the voice pattern data extraction unit 1111 is output by the voice recognition unit. The product category is extracted based on the voice pattern data (ST9-9). For example, when the voice pattern data of “anpan” and “anman” described above is stored in advance in the voice dictionary 1133 to be referred to and matches, the product category associated with the voice pattern data “anpan” and “anman” Two product categories “Anpan” and “Anman” are extracted as candidates by the voice pattern data extraction unit 1111.

  Next, the plurality of product category candidates extracted in ST9-9 are temporarily stored in the product category candidate storage unit 1141 (ST9-10). Here, product category candidates such as “Anpan” and “Anman” extracted in the category candidate area 81 of the product category candidate storage means 1141 of FIG. 8 are temporarily stored. Based on the candidates temporarily stored in the product category candidate storage unit 1141, the product reading unit 1113 reads the product data set corresponding to the product category from the product data / product category dictionary 1131 (ST9-11). ). Further, the CPU 111 obtains product sales data processing count information necessary when ordering the product data (ST9-12).

  Next, the output display control means 1161 ranks the product data based on the product sales data processing count information and rearranges it in a desired order (ST9-13). For example, a form in which product data with a large number of sales processes is preferentially displayed in order from the top is conceivable. In this case, since the items are displayed in order of product data with the largest number of sales data processing times, the items are displayed in descending order of possibility of performing the sales processing, and can be operated efficiently. In this way, the product data is appropriately rearranged in a desired order, and then displayed on the display 121 (ST9-14). Note that the information for order control by the output display control means 1161 is not limited to product sales data processing count information. For example, the order of the product name in alphabetical order, product price order, popular product order, recommended product, etc. It may be in order. As shown in FIG. 4, the product data is displayed on the display 36 at the center. As a display method of the product data at this time, the number of product data to be output and displayed in advance may be limited and displayed, or all may be displayed at once and displayed in a scrollable manner so as to be selectable. Further, not only the name of the product but also the sales information such as the image or price of the product data and the number of sales points may be displayed simultaneously with the product name 73.

  Next, the operator selects a desired item from the product data displayed on the display 121 (ST9-15). After the merchandise data is selected, the CPU 111 sells the merchandise data (ST9-16), and the sales process is terminated by the above operation (ST9-17).

  As described above, according to the present embodiment, when a product category associated with the product name or the appearance of the product is input by voice, a plurality of product category candidates similar in terms of speech are extracted. This product category candidate is set corresponding to each product data, and since a plurality of product category candidates are extracted, it is possible to prevent omission of product data extraction even when there is a misrecognition during voice recognition. . In addition, since the sales process can be performed without speaking the official product name, the burden on the operator is reduced. Furthermore, when output is displayed on the display 121, the ordering control is performed based on the product sales data processing count information and is displayed in a desired form for the operator. Therefore, even a person unfamiliar with the sales processing operation or a beginner can easily perform it. The effect which can perform sales processing operation is produced. In addition, since the desired product candidate is output and displayed on the product sales processing screen 38 only by voice input, the number of touch operations associated with the sales processing can be reduced, and erroneous operation can be prevented, and the input work time can be reduced. Since it can be shortened, it has the effect of performing sales processing quickly.

  Next, as a second embodiment, a case where a search based on matching or similarity of product character strings using speech recognition is applied to a product sales data processing apparatus will be described with reference to FIGS. The same constituent elements as those in the first embodiment are denoted by the same reference numerals, and detailed description thereof is omitted.

  First, in the speech dictionary 1133 shown in FIG. 6 of the second embodiment, standard speech pattern data obtained by quantifying the speech feature amount when a predetermined word is uttered is stored in advance. This voice pattern data is voice data for a character string of each product name, and this voice dictionary 1133 is used to recognize what product character string the input voice is.

  The product data / product character string dictionary 1132 is storage means for storing product data and product character strings of product names as product data stored in the voice dictionary 1133 in association with each other. The data contents of the product data / product character string dictionary 1132 will be described with reference to FIG. In this table, product data 1002 and a product character string 1001 set corresponding to the product data 1002 are stored. The product character string here is a character string that constitutes part or all of the product name as product data. As data stored in the dictionary, a character string of a product name is stored in the product character string area 1001, and then a product code 1003 assigned and set to each product data 1002 is stored in the menu code area. The product name 73 that is voice-recognized and output and displayed on the display 121 is set in the product name area. Here, a product character string of product name 73 such as “shoko”, “shoku”, “shokupa”, “shokupan”, “shokupanha” is set for the product name 73 of “8 breads”. For example, when the operator inputs “8 sheets of bread” by voice and is recognized as “shokupan hachi”, “8 sheets of bread” corresponds to the candidate. In addition, when “shokupan” is recognized, in addition to the product name 73 of “8 breads”, the product name “six breads” set corresponding to the same product string “shokupan” is also provided. Applicable as a candidate.

  The voice pattern data extraction unit 1111 refers to the voice pattern data of the product character string stored in the voice dictionary 1133 based on the voice pattern data output by the voice recognition unit, and extracts the product character string candidate of the product character string. Means. This voice pattern data extraction means 1111 does not extract only the product character string candidates having the voice pattern data of the product character string matching the front, but the product character string having the voice pattern data of the product character string matching the middle and the rear. You may extract about a candidate. Next, there is a product character string candidate storage unit 1142 as a unit for storing the product character string candidates extracted by the voice pattern data extraction unit 1111.

  Next, FIG. 11 shows the product character string candidates stored in the product character string candidate storage unit 1142. The product character string extracted by the voice pattern data extraction unit 1111 is stored in the product character string candidate area 1101 of the product character string candidate storage unit 1142. Subsequently, based on the product character string stored in the product character string candidate storage unit 1142, the product data reading unit 1113 reads a product character string candidate set corresponding to the product character string. This storage means may be provided in the HDD 114 like other dictionaries and storage means, or may be provided in the server.

  The operation process of the second embodiment will be described with reference to FIG. In the second embodiment, the same processes as ST9-1 to ST9-7 in the first embodiment are performed from ST12-1 to ST12-7. When the power is turned on, the merchandise sales data processing apparatus is activated, and when the CPU 111 displays the merchandise sales processing screen 38, the merchandise sales processing screen 38 using voice recognition shown in FIG. 3 is displayed (ST12-1). In this state of the screen, the start of the capturing of ambient noise from the microphone 13 and the start / end of the speech uttered by the operator by the on / off operation of the speech recognition key 37 are started (ST12-2). The captured audio is A / D converted from analog audio data to digital audio data by the A / D converter 118 (ST12-3). The digital audio data after A / D conversion is subjected to noise processing by the SS method (ST12-4). Thereafter, the CPU 111 causes the speech recognition engine 119 to execute feature extraction of the captured voice pattern data (ST12-5). Here, since the voice dictionary 1133 stores the product name and the voice pattern data of the product character string in association with each other, the operator performs voice input for all of the product names in relation to the import of the voice data in ST12-2. Alternatively, it may be performed as a part of the product name.

Next, the voice feature quantity of the input voice is compared with the voice feature quantity created in advance to determine whether they match (ST12-6). If it does not match or resembles a previously created speech feature (NO in ST12-6), speech recognition ends (ST12-17), and if it matches or resembles a speech feature (YES in ST12-6) Or, referring to the inside of the acoustic dictionary based on the similar voice feature quantity, the voice pattern data stored in association with the voice feature quantity is output (ST12-7).
Next, the voice dictionary 1133 is referred to based on the output voice pattern data, and it is determined whether there is a product character string stored in association with the voice pattern data (ST12-8). If there is no product character string stored in association with the output voice pattern data (NO in ST12-8), the voice recognition is terminated (ST12-17). Alternatively, the operator may be warned such as displaying an error message with the content “Please input with another product character string” or issuing a warning sound. On the other hand, when there is a commodity character string candidate stored in association with the output voice pattern data (YES in ST12-8), the voice pattern data extraction unit 1111 is based on the voice pattern data output by the voice recognition unit. The product character string is extracted (ST12-9). For example, if the voice input by the operator is “bread” and the voice pattern data output as matched or similar by the voice recognition engine 119 is “shokupan”, “shoku”, “choco”, these voice pattern data A product character string such as “Shokupan”, “Shoku”, and “Choko” stored in association with is extracted by the voice pattern data extracting means 1111.

  Next, the product character string candidates extracted in ST12-9 are temporarily stored in the product character string candidate storage unit 1142 (ST12-10). At this time, the product character string candidates are stored in the product character string candidate storage area 1101 of FIG. Based on the temporarily stored product character string candidates, the product reading unit 1113 reads the product data set corresponding to the product character string candidates from the product data / product character string storage unit 1132 (ST12-11). . Further, product sales data processing count information relating to the read product data is acquired (ST12-12). Here, the merchandise sales information may be related to the number of merchandise sales data processing.

  Next, the CPU 111 rearranges the product data in a desired order using the output display control means 1161 based on the product sales data processing count information (ST12-13). For example, a form in which product data with a large number of sales processes is preferentially displayed in order from the top can be considered. In this way, the product data is appropriately rearranged in a desired order, and then displayed on the display 121 (ST12-14). As a display method at this time, the number of product data to be output and displayed in advance may be limited and displayed, or all product names may be displayed at a time and displayed in a scrollable form so as to be selectable. Further, not only the name of the product but also the sales information such as the image or price of the product data and the sales number may be displayed at the same time.

  Next, the operator selects desired product data from the product data displayed on the display 121 (ST12-15). After the selection, the CPU 111 sells the product data (ST12-16), and the selling process is terminated by the above operation (ST12-17). Note that the category search function of the first embodiment and the product character string search function of the second embodiment do not need to be performed separately, and may be performed by combining these two embodiments.

  According to the second embodiment of the present invention, since the product data is output on the product sales processing screen 38 by inputting a part of the product name by voice, the operation is performed even when the product name 73 is not accurately remembered. It can be performed. As described above, since the merchandise sales process is performed by voice recognition, even an operator unfamiliar with the process operation can easily perform the process. Further, when a part or all of the product name is inputted by voice, even if the product name is erroneously recognized, the product data similar to the voice is output, so that the omission of product data can be prevented. In addition, because product data can be extracted and output by forward matching, middle matching, and backward matching of product character strings, the operator can perform sales processing even if the product name is in a state of remembering. Since it is not necessary to input the product name correctly, the burden associated with the sales process can be reduced.

  As a third embodiment, a product data input device using voice recognition will be described. In addition, the same number is attached | subjected to the same component as 1st, 2 embodiment, and detailed description is abbreviate | omitted.

  The input processing means 1112 is a means for performing input processing when arbitrary product data is selected from product data output and displayed by voice input. A means for storing the number of times the product data is input by the input processing means 1112 is an input number information storage means 1144. The merchandise sales data processing apparatus of the first embodiment executes processes such as input / registration, but the merchandise data input apparatus of the third embodiment is an apparatus that executes only the input process. Different from the embodiment.

  Next, a flowchart of the product data input process will be described with reference to FIG. The processing from the speech input processing corresponding to ST13-1 to ST13-5 of this embodiment to the extraction of speech candidates is the same as the processing of ST9-1 to ST9-5 of the first embodiment. Further, the process of outputting / displaying the product data on the display 121 based on the product category candidates extracted from the process of extracting the product category candidates from ST13-6 to ST13-11 is the first embodiment ST9-6. To ST9-11. In this embodiment, the product category is input by voice, and the acoustic dictionary 1134 is referred to based on the voice feature amount of the input voice. Then, the voice recognition means outputs the voice pattern data that matches or is similar. Based on the voice pattern data, the voice dictionary 1133 is referred to, and the voice pattern data extraction unit 1111 extracts the product category. Product data set corresponding to the extracted product category candidates is read from the product category / product data dictionary 1131 by the product data reading means 1113 (ST13-1 to ST13-12). The read product data is displayed on the display 121. Next, the operator performs selection input for the displayed product data, and the input processing means 1112 performs input processing (ST13-13), and the input processing is terminated (ST13-14).

  According to the third embodiment of the present invention, when a product category is input by voice, a plurality of similar product category candidates are extracted. By extracting a plurality of product category candidates in this way, there is an effect that product data set corresponding to the product category candidates can be displayed without omission. Further, when the product data is output and displayed on the display 121, it is performed in a state where the order data is appropriately controlled, so that the operator can perform input processing quickly. In addition, since product data that is highly likely to be subjected to input processing is preferentially displayed as an effect of other sequential control, the number of input processing operations can be reduced and input errors associated with input processing can be prevented.

  Next, a fourth embodiment will be described. In addition, the same number is attached | subjected to the same component as 1st, 2nd, 3rd embodiment, and detailed description is abbreviate | omitted. In the fourth embodiment, a part of a product name is input by voice, and a product character string candidate that matches or is similar to the product character string of the product name is extracted. This is a product data input device that extracts product data based on the extracted product character string candidates and outputs and displays it.

  The process of the fourth embodiment will be described with reference to FIG. ST14-1 to ST14-11 perform the same processing as in the second embodiment. In this embodiment, a product name is inputted by voice, and product character string candidates that match or are similar to the product character string voice data of the product name are extracted. The product data set corresponding to the product character string candidate is read by the product data reading means 1113 (ST14-12). The read product data is displayed on the display 121, and the operator performs selection input for the displayed product data, and the input processing means 1112 performs input processing (ST14-13).

  According to the present embodiment, by inputting a part of the product name by voice, the product data candidate is output on the display 121 based on the character string candidate of the product name. For this reason, even if the input process is unfamiliar and an operator who does not store the official product name operates the apparatus, the input process can be performed quickly and easily. In addition, since product data that is highly likely to be input is displayed preferentially, it is possible to prevent an input error associated with the input process. Due to such an effect, it is possible to support an unfamiliar operator's input operation.

  The present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying constituent elements without departing from the scope of the invention in the implementation stage, and is disclosed in the embodiment. Various inventions can be modified by appropriately combining a plurality of components.

1 is an external perspective view of a merchandise sales data processing apparatus according to an embodiment of the present invention. The block diagram which shows the principal part structure of the merchandise sales data processing apparatus concerning this invention. The figure of the processing screen of the goods sales data processing device concerning the present invention. The figure of the screen where the goods category candidate was output by the voice recognition means of the goods sales data processing apparatus concerning this invention. The figure which shows the data content preserve | saved at the acoustic dictionary of the merchandise sales data processing apparatus concerning this invention The figure which shows the data content preserve | saved at the audio | voice dictionary in the merchandise sales data processing apparatus concerning this invention. The figure which shows the data content preserve | saved at the merchandise data and merchandise category dictionary of the merchandise sales data processing apparatus concerning this invention. The figure which shows the data content preserve | saved at the merchandise category candidate memory | storage means of the merchandise sales data processing apparatus concerning this invention. The flowchart which shows the process of 1st Embodiment in the merchandise sales data processing apparatus concerning this invention. The figure which shows the data content preserve | saved at the merchandise data and merchandise character string dictionary of the merchandise sales data processing apparatus concerning this invention. The figure which shows the data content preserve | saved at the goods character string candidate of the goods sales data processing apparatus concerning this invention. The flowchart which shows the process of 2nd Embodiment in the merchandise sales data processing apparatus concerning this invention. The flowchart which shows the process of 3rd Embodiment in the merchandise sales data processing apparatus concerning this invention. The flowchart which shows the process of 4th Embodiment in the merchandise sales data processing apparatus concerning this invention.

Explanation of symbols

1111 Voice pattern data extracting unit 1112 Input processing unit 1113 Product data reading unit 1114 Product sales data processing unit 1131 Product category / product data dictionary 1132 Product data / product string dictionary 1133 Speech dictionary 1134 Acoustic dictionary 1141 Product category candidate storage unit 1142 Product Character string candidate storage means 1143 Product sales data processing count information storage means 1144 Input count information storage means
12 Output display means

Claims (12)

An acoustic dictionary that stores voice feature data and voice pattern data created in advance in association with each other;
A voice dictionary storing the voice pattern data and product categories in association with each other;
Product data for identifying products, and a product category / product data dictionary storing the product categories set corresponding to the product data;
Voice input means for inputting voice;
The speech pattern stored in association with the matched or similar speech feature amount by comparing the speech feature amount of the speech input by the speech input unit with a previously created speech feature amount with reference to the acoustic dictionary Voice recognition means for outputting data;
Voice pattern data extracting means for referring to the voice dictionary and extracting the product category as a candidate based on the voice pattern data output by the voice recognition means;
Commodity category candidate storage means for storing the commodity category extracted by the voice pattern data extraction means as commodity category candidates;
Product data reading means for reading the product data set corresponding to the product category candidate from the product category / product data dictionary;
Output display means for outputting and displaying the product data read by the product data reading means;
Product sales data processing means for performing product sales data processing when any of the product data is selected from the product data output and displayed by the output display means;
A product sales data processing apparatus characterized by comprising: An acoustic dictionary that stores voice feature data and voice pattern data created in advance in association with each other;
A voice dictionary storing the voice pattern data and product character strings in association with each other;
Product data for identifying the product and product data / product character string dictionary storing the product character string set corresponding to the product data;
Voice input means for inputting voice;
The speech pattern stored in association with the matched or similar speech feature amount by comparing the speech feature amount of the speech input by the speech input unit with a previously created speech feature amount with reference to the acoustic dictionary Voice recognition means for outputting data;
Voice pattern data extracting means for referring to the voice dictionary and extracting the product character string as a candidate based on the voice pattern data output by the voice recognition means;
Commodity character string candidate storage means for storing the commodity character string extracted by the voice pattern data extraction means as a commodity character string candidate;
Commodity data reading means for reading out the commodity data set corresponding to the commodity character string candidate from the commodity data / commodity character string dictionary;
Output display means for outputting and displaying the product data read by the product data reading means;
Product sales data processing means for performing product sales data processing when any of the product data is selected from the product data output and displayed by the output display means;
A product sales data processing device comprising: The product sales data processing device includes: output display control means for performing output control and displaying the product data read by the product data reading means;
The merchandise sales data processing apparatus according to claim 1 or 2, further comprising: The product sales data processing means includes product sales data processing frequency information storage means for storing the number of times product sales data processing has been performed on the product data.
The output display control means controls the order of the product data based on the product sales data processing count information stored in the product sales data processing count information storage means, and outputs and displays it.
The merchandise sales data processing apparatus according to claim 3. An acoustic dictionary that stores voice feature data and voice pattern data created in advance in association with each other;
A voice dictionary storing the voice pattern data and product categories in association with each other;
Product data for identifying products, and a product category / product data dictionary storing the product categories set corresponding to the product data;
Voice input means for inputting voice;
The speech pattern stored in association with the matched or similar speech feature amount by comparing the speech feature amount of the speech input by the speech input unit with a previously created speech feature amount with reference to the acoustic dictionary Voice recognition means for outputting data;
Voice pattern data extracting means for referring to the voice dictionary and extracting the product category as a candidate based on the voice pattern data output by the voice recognition means;
Merchandise category candidate storage means for storing the quotient 74. product category extracted by the voice pattern data extraction means as a merchandise category candidate;
Product data reading means for reading the product data set corresponding to the product category candidate from the product category / product data dictionary;
Output display means for outputting and displaying the product data read by the product data reading means;
When any product data is selected from the product data output and displayed by the output display means, input processing means for performing input processing;
A product data input device comprising: An acoustic dictionary that stores voice feature data and voice pattern data created in advance in association with each other;
A voice dictionary storing the voice pattern data and product character strings in association with each other;
Product data for identifying the product and product data / product character string dictionary storing the product character string set corresponding to the product data;
Voice input means for inputting voice;
The speech pattern stored in association with the matched or similar speech feature amount by comparing the speech feature amount of the speech input by the speech input unit with a previously created speech feature amount with reference to the acoustic dictionary Voice recognition means for outputting data;
Voice pattern data extracting means for referring to the voice dictionary and extracting the product character string as a candidate based on the voice pattern data output by the voice recognition means;
Commodity character string candidate storage means for storing the commodity character string extracted by the voice pattern data extracting means as a commodity character string candidate;
Commodity data reading means for reading out the commodity data set corresponding to the commodity character string candidate from the commodity data / commodity character string dictionary;
Output display means for outputting and displaying the product data read by the product data reading means;
When any product data is selected from the plurality of product data output by the output display means, input processing means for performing input processing;
A product data input device comprising: The product data input device includes output display control means for performing output display by orderly controlling the product data read by the product data reading means,
The product data input device according to claim 1, further comprising: The input data processing means has input number information storage means for storing the number of times of input of the selected product data. The product data is based on the input number information stored in the input number storage means by the output display control means. Display and control the order of
The product data input device according to claim 3. Corresponding to an acoustic dictionary that stores voice feature values and voice pattern data that are created in advance in association with each other, a voice dictionary that stores the voice pattern data in association with product categories, product data that identifies products, and this product data In the product sales data processing apparatus provided with the product category / product data dictionary storing the product category set as described above,
Voice input function to input voice,
The speech pattern stored in association with the matched or similar speech feature amount by comparing the speech feature amount of the speech input by the speech input unit with a previously created speech feature amount with reference to the acoustic dictionary Voice recognition function to output data,
A product category extraction function for referring to the voice dictionary based on the output voice pattern data and extracting the product category as a candidate;
A product category storage function for storing the product categories extracted as candidates;
A product data reading function for reading the product data set corresponding to the stored product category from the product category / product data dictionary;
An output display function for outputting and displaying the read product data;
A product sales data processing function for processing product sales data when any of the product data is selected from the output product data;
A program to realize An acoustic dictionary in which voice feature values of voice created in advance and voice pattern data are stored in association with each other, a voice dictionary in which the voice pattern data and product character strings are stored in association with each other, product data for identifying products, and the product data In the product sales data processing apparatus provided with the product data / product character string dictionary storing the product character string set correspondingly,
Voice input function to input voice,
The speech pattern stored in association with the matched or similar speech feature amount by comparing the speech feature amount of the speech input by the speech input unit with a previously created speech feature amount with reference to the acoustic dictionary Voice recognition function to output data,
A product character string extraction function for referring to the voice dictionary based on the output voice pattern data and extracting the product character string as a candidate;
A product character string storage function for storing the product character string extracted as a candidate;
A product data read function for reading the product data set corresponding to the stored product character string from the product data / product character string dictionary;
An output display function for outputting and displaying the read product data;
A product sales data processing function for processing product sales data when any of the product data is selected from the plurality of product data output and displayed;
A program to realize Corresponding to an acoustic dictionary that stores voice feature values and voice pattern data that are created in advance in association with each other, a voice dictionary that stores the voice pattern data in association with product categories, product data that identifies products, and this product data In the product data input device provided with the product category / product data dictionary storing the product category set as described above,
Voice input function to input voice,
Voice pattern data stored in association with a matched or similar voice feature quantity by comparing the voice feature quantity of the voice inputted by the voice input means with a previously created voice feature quantity with reference to the acoustic dictionary Voice recognition function that outputs
A product category extraction function for referring to the voice dictionary based on the output voice pattern data and extracting the product category as a candidate;
A product category storage function for storing the product categories extracted as candidates;
A product data reading function for reading the product data set corresponding to the stored product category from the product category / product data dictionary;
An output display function for outputting and displaying the read product data;
When any product data is selected from the product data output and displayed, an input processing function for performing input processing;
A program to realize An acoustic dictionary in which voice feature values of voice created in advance and voice pattern data are stored in association with each other, a voice dictionary in which the voice pattern data and product character strings are stored in association with each other, product data for identifying products, and the product data In the product data input device provided with the product data / product character string dictionary storing the product character string set correspondingly,
Voice input function to input voice,
Refer to the acoustic dictionary, compare the voice feature quantity of the voice input by the voice input means with a voice feature quantity created in advance, and store voice pattern data associated with the matched or similar voice feature quantity Voice recognition function that outputs
A product character string extraction function for referring to the voice dictionary based on the output voice pattern data and extracting the product character string as a candidate;
A product character string storage function for storing the product character string extracted as a candidate;
A product data read function for reading the product data set corresponding to the stored product character string from the product data / product character string dictionary;
An output display function for outputting and displaying the read product data;
An input processing function for performing input processing when any of the product data is selected from the plurality of output product data;
A program to realize

Images