Information processing device, control method, and program

The mechanism dynamically adjusts the speech recognition dictionary based on data type and confidence levels to improve accuracy in recognizing specific words, addressing partial optimization challenges.

JP7587104B2Active Publication Date: 2024-11-20CANON MARKETING JAPAN INC +1
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
JP2020115311
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2020-07-03
Publication Date
2024-11-20
Estimated Expiration
2040-07-03

AI Technical Summary

Technical Problem

Existing speech recognition technologies struggle with partial optimization, failing to accurately recognize specific words like proper nouns and homonyms, leading to suboptimal recognition results.

Method used

A mechanism that dynamically adjusts the word dictionary based on data type and confidence levels, using management means to specify data types, recognition execution, and value obtaining means to enhance accuracy.

Benefits of technology

Enhances speech recognition accuracy by dynamically adapting the word dictionary, ensuring optimal recognition of specific data items even in challenging conditions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007587104000001
    Figure 0007587104000001
  • Figure 0007587104000002
    Figure 0007587104000002
  • Figure 0007587104000003
    Figure 0007587104000003
Patent Text Reader

Abstract

To provide a mechanism that determines a word to be used for voice recognition in accordance with contents of voice data, thereby acquiring an appropriate voice recognition result.SOLUTION: In an information processing system, an information processing server 100 is connected to a recognition server 101 that performs voice recognition on voice data so as to communicate with each other, and includes a recognition pattern determination unit 302 that, in accordance with a data item recognized from the voice data, determines a word related to the data item to be used for the voice recognition, and a pattern processing unit 304 that acquires a value of the data item based on a result of the voice recognition performed by the voice recognition device by using the determined word.SELECTED DRAWING: Figure 3
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to a voice recognition technology for converting voice data into a character string. [Background technology]

[0002] Speech recognition technology has been used for the purpose of improving the efficiency of input operations into information processing devices, etc. In recent years, the accuracy of speech recognition has improved considerably thanks to technologies such as machine learning.

[0003] Cloud services and the like have humans analyze large amounts of data and use it as training data for machine learning to further improve accuracy, but users still need to customize it to suit their needs.

[0004] Words may also be registered depending on the purpose. In particular, proper nouns such as unusual personal names, department names unique to an organization, and product names are not included in general dictionaries, and may be registered by matching them with pronunciation. In addition, in cases where a term is common but has many homonyms, it is not necessarily necessary to match it with pronunciation, but it may be registered in order to raise its priority.

[0005] There are several methods for registering words in a voice recognition service.

[0006] In one service, necessary words are registered in advance and a series of voice recognition requests are repeated. In this method, the same word information registered in advance is used for each voice recognition request, so even if there are homophones with different notations, they may not be used appropriately.

[0007] In another example of a service, there is a method in which the words to be registered can be changed according to the contents of the voice data to be passed to the service each time a voice recognition request is made. In this method, the number of words that can be registered is limited to a relatively small number, so appropriate words must be selected for each voice recognition request.

[0008] The technology described in Patent Document 1 involves trial-registering words that are candidates for registration, performing voice recognition on evaluation voice data, and then presenting the effect to a human. If the human determines that the word should be registered, it is added to the word dictionary. [Prior art documents] [Patent documents]

[0009] [Patent Document 1] JP 2018-40906 A Summary of the Invention [Problem to be solved by the invention]

[0010] The technology of Patent Document 1 registers all words that are determined to be registered in advance. However, in this method, the registered words affect any voice data, so while it is highly likely to be effective in improving the overall accuracy of voice recognition, it may not return appropriate results for individual voice recognition requests. In other words, instead of achieving global optimization, it may not be locally optimal.

[0011] The object of the present invention is to Identify the sound Voice recognition results to Value of data item The purpose of this study is to provide a mechanism for obtaining such information. [Means for solving the problem]

[0012] The present invention includes a management means for managing the format of a data item corresponding to a data type, a recognition execution means for acquiring first text data obtained by speech recognition from speech data using a first word list, and a recognition execution means for recognizing the speech data based on the acquired first text data. of A type identification means for identifying a data type, and a value acquisition means for acquiring a value of a data item corresponding to the identified data type from the first text data based on a format of the data item. The recognition execution means acquires second text data that has been speech-recognized from the speech data using a second word list related to the specified data type, and the value acquisition means acquires a value of a data item corresponding to the data type from the second text data based on a format of the data item. It is characterized by: Effect of the Invention

[0013] According to the present invention, the contents of the audio data Identify the sound Voice recognition results to Value of data item It is possible to provide a mechanism for obtaining the above information. [Brief description of the drawings]

[0014] [Figure 1] 1 is a diagram illustrating an example of a configuration of an information processing system according to an embodiment of the present invention. [Diagram 2] 1 is a block diagram showing an example of a hardware configuration applicable to an information processing server 100, a recognition server 101, and an information processing terminal 102 according to an embodiment of the present invention. [Diagram 3] FIG. 2 is a diagram illustrating an example of a functional configuration according to an embodiment of the present invention. [Figure 4] 2 is a diagram showing an example of a user interface of the information processing terminal 102 according to the embodiment of the present invention. [Diagram 5] 2 is a diagram showing an example of a user interface of the information processing terminal 102 according to the embodiment of the present invention. [Figure 6] 4 is a flowchart illustrating an example of a voice recognition process according to an embodiment of the present invention. [Figure 7] 1 is an example of a flowchart illustrating a process for determining a word to be registered according to an embodiment of the present invention. [Figure 8] 3 is a diagram illustrating an example of a storage unit that stores items to be speech-recognized and related information in association with each other according to the embodiment of the present invention; FIG. [Figure 9] 3A and 3B are diagrams illustrating an example of voice data input by a user and a voice recognition result according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

[0016] FIG. 1 is a diagram showing an example of a configuration of an information processing system according to an embodiment of the present invention.

[0017] The information processing terminal 102, the information processing server 100, and the recognition server 101 are communicably connected via a network 103. The network 103 is, for example, the Internet or a LAN.

[0018] The information processing terminal 102 is a personal computer or a smartphone on which an application that constitutes the user interface in the present invention is installed, or on which web browser software for executing a web application provided by the information processing server 100 is installed, and the web application is accessed by specifying a specific URL.

[0019] The information processing terminal 102 also has a built-in microphone for inputting voice data in the present invention or is connected to the information processing terminal 102 as an external device. A camera used for a purpose described later may also be built-in or connected to the information processing terminal 102 as an external device.

[0020] The information processing server 100 provides the application of the present invention to the information processing terminal 102, and sends voice data received from the information processing terminal 102 to the recognition server 101, which executes voice recognition processing and receives the results (text data).

[0021] However, the information processing server 100 and the recognition server 101 may be in the same housing. Also, some of the functions of the application of the present invention may be arranged and executed in the information processing terminal 102. The arrangement of functions described below is merely an example, and the arrangement is not limited to a specific information processing device. This concludes the explanation of FIG. 1.

[0022] FIG. 2 is a block diagram showing an example of a hardware configuration applicable to the information processing server 100, the recognition server 101, and the information processing terminal 102 according to the embodiment of the present invention.

[0023] As shown in FIG. 2, the information processing server 100, the recognition server 101, and the information processing terminal 102 are configured in such a manner that a CPU (Central Processing Unit) 201, a RAM (Random Access Memory) 203, a ROM (Read Only Memory) 202, an input controller 205, a video controller 206, a memory controller 207, a communication I / F controller 208, etc. are connected via a system bus 204.

[0024] The CPU 201 comprehensively controls each device and controller connected to a system bus 204 .

[0025] The ROM 202 or the external memory 211 also stores a BIOS (Basic Input / Output System) and an OS (Operating System), which are control programs for the CPU 201, and various programs (described later) required to realize the functions executed by each server or each PC. Information required to implement the present invention is also stored. The external memory may be a database.

[0026] The RAM 203 functions as a main memory, a work area, etc., for the CPU 201. The CPU 201 loads programs and the like required for executing processes from the ROM 202 or the external memory 211 into the RAM 203, and executes the loaded programs to realize various operations.

[0027] In addition, the input controller 205 controls input from a keyboard (KB) 209 and a pointing device such as a mouse (not shown).

[0028] A video controller 206 controls the display on a display device such as a display 210. The display device may be a liquid crystal display or the like. These are used by the administrator as necessary.

[0029] The memory controller 207 controls access to an external memory 211 such as an external storage device (hard disk (HD)) that stores boot programs, various applications, font data, user files, edited files, various data, etc., a flexible disk (FD), or a compact flash (registered trademark) memory connected to a PCMCIA (Personal Computer Memory Card International Association) card slot via an adapter.

[0030] A communication I / F controller 208 connects to and communicates with external devices via a network, and executes communication control processing on the network. For example, communication using TCP / IP (Transmission Control Protocol / Internet Protocol) is possible.

[0031] The CPU 201 can display on the display 210 by, for example, executing a process of expanding (rasterizing) an outline font in a display information area in the RAM 203. The CPU 201 also enables a user to give instructions using a mouse cursor (not shown) on the display 210.

[0032] Various programs for implementing the present invention, which will be described later, are recorded in the external memory 211, and are executed by the CPU 201 by being loaded into the RAM 203 as required.

[0033] 3 is a diagram showing an example of a functional configuration according to an embodiment of the present invention. A data input unit 321 receives data such as voice by a user operating the application of the present invention. It may also receive images of receipts related to an item (in the example described later, "payment item").

[0034] The data transmission unit 322 transmits data such as voice data input by a user operation to the information processing server 100, and the data reception unit 301 of the information processing server 100 receives the data.

[0035] The recognition pattern determination unit 302 judges items that are not yet determined at that time based on the "items" in the data ("payment items" in the example described below) and already determined "detailed items" ("start date of stay" and "lodging fee" in the example described below), and determines the recognition method. The recognition method means, for example, determining which words to register and to perform voice recognition. Information required for these determinations is stored in the rule storage unit 306.

[0036] Based on the determined recognition method, the recognition execution unit 303 transmits the registered words and voice data to the recognition server 101, causes the voice recognition unit 311 to recognize the registered words and voice data, and receives the result (recognized character string). As will be described later, image data such as a receipt may be sent to the character recognition unit 312, which then performs recognition and receives the result (recognized character string).

[0037] The information processing server 100 identifies the parts of the character string recognized by the recognition server 101 that correspond to the "detailed items" (such as "stay start date" and "lodging fee") in the pattern processing unit 304. At that time, the information described in the pattern 804 stored in the rule storage unit 306 is used as a pattern matching rule.

[0038] The result transmission unit 305 transmits the above-mentioned specified information to the information processing terminal 102, and the result reception unit 324 of the information processing terminal 102 receives it. This information is presented to the user on the display unit 323, and the user is prompted to perform the next operation (such as "re-recognition" if the information is insufficient, or "registration" if the information is sufficient). This completes the explanation of FIG. 3.

[0039] Next, an example of a screen for expense settlement will be described with reference to Figures 4 and 5. This screen shows an example of a user interface on the information processing terminal 102.

[0040] First, Fig. 4 will be explained. When the user launches the "expense settlement application" on the information processing terminal 102, a settlement item selection list 401 (list selection) is displayed for selecting the type of expense to be settled. If the user knows the detailed item to be settled and selects it from the list, the flow of the process explained in the flowchart below becomes simple, but if the user does not know, there is no need to make a selection.

[0041] Also, by pressing the voice recognition button 402 , for example, a voice message such as “I will stay at the Yokohama Hotel from July 10th to July 12th for a total of 19,800 yen” is received by the information processing terminal 102 and sent to the information processing server 100 .

[0042] FIG. 5 shows the result of recognizing the voice input by the user, indicating that the bill item is "accommodation fee," and extracting and displaying information about detailed items ("start date of stay," "end date of stay," "location," and "cost"). Column 501 is marked with "◎" or "×," which indicate whether the application was successful in obtaining the information. In this example, the name of the accommodation was not properly recognized ("×").

[0043] Whether the information was successfully acquired can be changed at the user's discretion. For example, in the example in Figure 5, "Cost" is marked with a "◎", but this is the application's judgment and may actually be incorrect.

[0044] The user may change these recognition results by inputting text, or may simply correct the "◎" or "X" in field 501 and then press the voice recognition button 402 again.

[0045] At that time, the expense report is for "accommodation," and only the information regarding "lodging" is inaccurate, and this information is sent to the information processing server 100. In the process flow described below, the application uses the information already known and the information designated as inaccurate to perform more appropriate voice recognition and information acquisition, and displays it on the screen again.

[0046] Finally, if all the results are entered correctly, the user presses the registration button 502 to claim the expenses.

[0047] Needless to say, these processes are not limited to "expense settlement" and can be used for various forms, etc. This completes the explanation of the screen examples using Figures 4 and 5.

[0048] Next, an example of processing in the present invention will be described with reference to the flowcharts of Figures 6 and 7. Each step of the flowchart is executed by the CPU 201 on the information processing server 100. In addition, the rule storage unit 306 (Figure 8) that is referenced when executing each step, and an example of voice data input by the user and a voice recognition result (Figure 9) will also be described as necessary.

[0049] In step S601, data transmitted from the information processing terminal 102 is received. In addition to the voice data uttered by the user, if any "item" is selected in the payment item selection list 401, the data also includes information about which item (e.g., "accommodation fee") is selected.

[0050] For example, assume that the user's initial speech input is "I will stay at the Yokohama Hotel from July 10th to July 12th for a total of 19,800 yen" (user's initial utterance 901 in FIG. 9).

[0051] In step S602, a recognition pattern is identified and a list of words to be registered in the speech recognition unit 311 of the recognition server 101 in order to optimize the speech recognition accuracy for that purpose is made. This process is explained with reference to the flowchart of FIG.

[0052] In step S701, it is determined whether the "item" (in this example, the "payment item") for which detailed information should be determined has been determined. For example, when voice data is sent from the information processing terminal 102 to the information processing server 100 for the first time and the user has not explicitly selected an item in the payment item selection list 401, it is necessary to determine it as a result of voice recognition (i.e., it has not been determined at this point). In this case (if NO), proceed to step S702.

[0053] If the user has explicitly selected the item in the payment item selection list 401, or if voice recognition has already been performed once and the payment item to be processed has been selected from the results (YES), the process proceeds to step S703.

[0054] In step S702, since the "payment item" is undecided, priority is given to improving the accuracy of speech recognition to determine the "payment item" rather than obtaining detailed items (such as "accommodation"). Therefore, the words described in payment item characteristic words 802 in FIG. 8 are listed as words to be registered in the speech recognition service (words included in characteristic word groups S1 to SZ). This process is an example of a process to determine words related to the type of speech data to be used in speech recognition.

[0055] We will now explain the words listed in payment item characteristic word 802 in Fig. 8. At this point, it is necessary to determine the "payment item" (for example, "accommodation fee" or "taxi fare") rather than detailed information on the payment content, and priority is given to accurately acquiring words (or character strings) for distinguishing between them from the voice data.

[0056] Therefore, in the payment item characteristic word 802, a word (or character string) such as that in the characteristic word group S1 is defined as a word that characterizes, for example, "accommodation fee". If the information included in the voice data is "accommodation fee", then it is considered that there is a high probability that, for example, "accommodation" or "hotel" will appear in the user's voice, so it is important to register these words and reliably obtain them as the results of voice recognition. Similarly, if it is "taxi fare", then there is a high probability that the voice will include "taxi", and if it is "train fare", then there is a high probability that the voice will include the name of the line, such as "Toei", "Metro", or more specifically, "Asakusa Line". These words are listed up for registration in the voice recognition unit 311 before a voice recognition request, which will be described later.

[0057] Here, the flow chart temporarily returns to step S603 in Fig. 6. In step S603, the words included in the feature word groups S1 to SZ determined in the above-mentioned step S702 are transmitted to the speech recognition unit 311 of the recognition server 101 for registration in order to improve the speech recognition accuracy, and the speech data (901) to be recognized is also transmitted.

[0058] In step S604, the result of the recognition (character string as a result of the recognition) is received by the voice recognition unit 311. For example, the "first voice recognition result" (902) in FIG.

[0059] In step S605, the character string of the recognition result received in step S604 is checked to see if it contains information corresponding to the "payment item" and "detailed item". Specifically, since the characters "hotel" and "lodging" are included in 902, it can be determined that the "payment item" is "accommodation fee". This process is an example of a process for identifying the type of voice data based on the result of speech recognition by a speech recognition device using the determined words. In addition, when determining the "payment item", it may be determined not only when the speech recognition result 902 contains a word that matches the feature word 806 as described above, but also when it contains a word similar to the feature word or a related word, and may be determined by pattern matching with the recognition result in which the "scrutiny item" was determined in the past, inference by machine learning, or the like. Furthermore, information on the "end date of stay" and "cost" can be obtained from the detailed items related to the payment item "accommodation fee". Specifically, this is because the character strings 922 and 925 in FIG. 9 match the pattern described in pattern 804 in FIG. 8. This process is an example of a process for extracting portions related to data items based on predetermined conditions from the results of speech recognition by a speech recognition device.

[0060] In step S606, it is confirmed whether or not all detailed items have been acquired for the settlement item in question. In the above example, since the "lodging start date" and "lodging location" of the "lodging fee" have not been acquired, the result is NO and the process proceeds to step S607. This process is an example of a process for determining whether or not the value of a data item has been acquired by the acquisition means.

[0061] If all detailed items have been acquired (if "YES"), the recognition result is returned to the information processing terminal 102, and the processing of the present invention is completed.

[0062] Next, steps S607 and S608 will be explained, but these judgments are merely examples. That is, the essence of the present invention is to obtain necessary information by dynamically switching the dictionary registered in the voice recognition unit 311 even with the same voice data, but if the quality of the initial voice data is poor, it will not be possible to obtain the information no matter how many times it is repeated. Therefore, a certain limit is imposed on the reuse of voice data. Since these are merely examples, it goes without saying that any other judgment criteria, such as the user's waiting time or the number of detailed items that could not be recognized after repeated attempts, can also be used to configure the present invention. Now, steps S607 and S608 will be explained as examples.

[0063] In step S607, the confidence level of the speech recognition is used as a criterion for judgment. As a general technology, the output of speech recognition includes not only the character string of the recognition result but also the "confidence level." For example, the closer to 1 on a numerical scale from 0 to 1, the higher the probability that the answer is correct (matches the user's speech).

[0064] When the confidence level is high (YES), the quality of the voice data input by the user is high to begin with, for example, the recording environment is quiet, the performance of the input device such as a microphone is good, and the speaker has good pronunciation, so even if the initial voice recognition fails, there is a high possibility that the correct answer will be output by registering words, etc. Therefore, the process returns to step S602 via step S608 and repeats the process again.

[0065] If the confidence level is low, the correct answer will not be obtained even if the word is registered and recognized again (if NO), so the process proceeds to step S609. The threshold value for the confidence level at which the correct answer can be obtained / not obtained is manually set in advance in a storage unit such as a file.

[0066] In step S608, a limit is set on the number of times the speech data is recognized. This makes it possible to avoid repeating a situation in which, for example, a word that is a characteristic word is registered in the speech recognition unit 311 and the confidence level is improved, but the necessary information is not actually obtained. This number of times is also manually set in advance in the storage unit such as the file. Alternatively, instead of the number of times, if the characteristic word group in FIG. 8 cannot be recognized even when it is expanded to the maximum, the answer may be determined as "YES" and the process may proceed to step S609.

[0067] If the result in step S608 is NO, the process returns to step S602 (i.e., the flowchart in FIG. 7) again, and preparations are made for re-recognizing the initial voice data.

[0068] 7 is executed for the second or subsequent time, the "payment item" has already been determined ("accommodation fee" in this example). In this case, "YES" is determined in step S701, and the process proceeds to step S703.

[0069] In step S703, preparations are made to pass words belonging to the characteristic word groups of the "detailed items" that have not yet been determined to the speech recognition unit 311. In this example, specifically, since "start date of stay" and "location" have not yet been acquired, words in the related characteristic word groups D1 and D2 are listed, and the process returns to the flowchart in Fig. 6. This process is an example of a process for determining words related to data items to be used in speech recognition according to the data items recognized from the speech data.

[0070] 6, the speech recognition unit 311 is instructed to register words in the feature word groups D1 and D2, and the speech data initially input by the user is passed to the speech recognition unit 311. This process is an example of a process for determining words related to a data item as words to be used in speech recognition according to the judgment result.

[0071] The result is received in step S604, and in step S605 it is confirmed whether these two detailed items can be obtained. The start date of the stay ("July, this department" (921)) failed to be recognized in 902 of Fig. 9, but it was correctly recognized the second time and "July 10th" was obtained (903). However, suppose that "Place of lodging" is "?? Hotel" (931), and it fails again. This process is an example of a process for obtaining the value of a data item based on the result of speech recognition by a speech recognition device using determined words.

[0072] Therefore, in steps S607 and S608, it is determined that the first voice data should be used to perform the third voice recognition.

[0073] Again, in step S703, since there is only one detailed item to be recognized, "accommodation", it is determined that more words can be registered than the second time, and the words in expanded characteristic word group D2 in expanded characteristic word table 810 in Fig. 8 are registered. By registering more hotel names than in characteristic word group D2, it is attempted to increase the probability of successful recognition. However, it ultimately fails as in 941.

[0074] As a result, it is determined that the same voice data should not be recognized again (if "YES" in step S608), and in step S609, an instruction is issued to the information processing terminal 102 to input only "accommodation" by voice again. This process is an example of a process for accepting input of voice data related to a data item that could not be acquired by the acquisition means.

[0075] When the user realizes that only the “accommodation” of the “accommodation fee” has not been correctly acquired on the information processing terminal 102 (501 in FIG. 5), he or she inputs only that information by voice and transmits it to the information processing server 100 again.

[0076] In step S601, the information processing server 100 receives the information, recognizes only "accommodation" using the same procedure as in the second and subsequent speech recognition attempts described above, and acquires the information. If the information acquisition is successful, the result is returned to the information processing terminal 102 as the final result (step S610). If the information cannot be acquired again, a failure message is returned in step S609, as in the previous period.

[0077] With the above, it is assumed that all the detailed items of "accommodation fee" have been input, and the explanation using the flowcharts in Fig. 6 and Fig. 7, the storage unit in Fig. 8, and the execution example in Fig. 9 is completed. Note that even if the system considers it successful, it may simply mean that the patterns matched, and the user may not have obtained accurate data. In such a case, it goes without saying that the user can select just that item again from the interface of the information processing terminal 102 and input voice data, or give up on inputting information by voice data and input characters directly.

[0078] 6 to 9 are merely examples, and since the focus of the present invention is to change what information (word group) is to be prioritized for acquisition at each stage, even if the same voice data is repeatedly acquired during the process of continuous voice recognition, the method of deciding the priority information to be acquired and the word group to be registered at each stage is arbitrary, and it goes without saying that all of these are included in the scope of the present invention. Furthermore, the same is true for the timing of re-requesting different voice data rather than the same voice data, and any method of controlling the flow is included in the present invention.

[0079] Although several embodiments have been described above, the present invention can be embodied, for example, as a system, an apparatus, a method, a computer program, or a recording medium, and specifically, the present invention may be applied to a system made up of multiple devices, or may be applied to an apparatus made up of a single device.

[0080] Furthermore, the computer program of the present invention is a computer program that enables a computer to execute the processing method of the flowcharts shown in Figures 6 and 7, and the storage medium of the present invention stores a computer program that enables a computer to execute the processing method of Figures 6 and 7. Note that the computer program of the present invention may be a computer program for each processing method of each device in Figures 6 and 7.

[0081] As described above, it goes without saying that the object of the present invention can be achieved by supplying a recording medium on which a computer program that realizes the functions of the above-mentioned embodiments is recorded to a system or device, and having the computer (or CPU or MPU) of that system or device read and execute the computer program stored on the recording medium.

[0082] In this case, the computer program itself read out from the recording medium will realize the novel functions of the present invention, and the recording medium storing the computer program will constitute the present invention.

[0083] Examples of recording media for supplying a computer program include flexible disks, hard disks, optical disks, magneto-optical disks, CD-ROMs, CD-Rs, DVD-ROMs, magnetic tapes, non-volatile memory cards, ROMs, silicon disks, solid state drives, etc.

[0084] Furthermore, it goes without saying that the functions of the above-mentioned embodiments are not only realized by the computer executing a computer program read by the computer, but also include cases where an OS (operating system) running on the computer performs some or all of the actual processing based on the instructions of the computer program, and the functions of the above-mentioned embodiments are realized through that processing.

[0085] Furthermore, it goes without saying that this also includes cases where a computer program read from a recording medium is written into a memory provided on an expansion board inserted into a computer or a function expansion unit connected to the computer, and then a CPU or the like provided on the expansion board or function expansion unit performs some or all of the actual processing based on the instructions of the computer program code, thereby realizing the functions of the above-mentioned embodiments.

[0086] The present invention may be applied to a system consisting of multiple devices, or to an apparatus consisting of a single device. Needless to say, the present invention can also be applied to a case where the present invention is achieved by supplying a computer program to a system or apparatus. In this case, the system or apparatus can enjoy the effects of the present invention by reading a recording medium storing a computer program for achieving the present invention into the system or apparatus.

[0087] Furthermore, by downloading and reading out a computer program for achieving the present invention from a server, database, etc. on a network using a communication program, the system or device can enjoy the effects of the present invention.

[0088] It should be noted that the present invention also includes configurations that combine the above-described embodiments and their modified examples. [Explanation of symbols]

[0089] 100 Data Processing Server 101 Recognition Server 102 Information processing terminal 103 Network 301 Data Reception Unit 302 Recognition pattern determination unit 303 Recognition Execution Unit 304 Pattern processing section 305 Result transmission unit 306 Rule Memory Unit 311 Voice Recognition Unit 312 Character recognition section 321 Data Entry Department 322 Data transmission unit 323 Display section 324 Result Reception Unit

Claims

1. A management means for managing a format of a data item corresponding to a data type; a recognition execution means for acquiring first text data that is speech-recognized from the speech data using the first word list; a type identification means for identifying a data type of the voice data based on the acquired first text data; a value acquiring means for acquiring a value of a data item corresponding to the specified data type from the first text data based on a format of the data item; Equipped with the recognition execution means acquires second text data obtained by speech recognition from the speech data using a second word list related to the specified data type; The value acquiring means acquires a value of a data item corresponding to the data type from the second text data based on a format of the data item.

23. An information processing apparatus comprising:

2. 2 . The information processing apparatus according to claim 1 , wherein the second word list includes words related to a data item corresponding to the data type.

3. 3. The information processing apparatus according to claim 1, further comprising a second generating means for generating the second word list.

4. the second generation means generates a second word list related to a data item for which a value could not be acquired by the value acquisition means, among data items corresponding to the data type; The recognition execution means acquires second text data that has been speech-recognized from the speech data using the generated second word list; 4. The information processing apparatus according to claim 3, wherein the value acquisition means reacquires the value of the data item that could not be acquired from the second text data.

5. a receiving means for receiving input of voice data relating to a data item whose value cannot be acquired by the value acquiring means, the recognition execution means acquires third text data that is speech-recognized from the input speech data using a second word list related to the data item that could not be acquired; 5. The information processing apparatus according to claim 1, wherein the value acquisition means reacquires the value of the data item that could not be acquired from the third text data.

6. 6. The information processing apparatus according to claim 1, wherein the first word list includes words capable of identifying a data type of the voice data.

7. 7. The information processing apparatus according to claim 1, further comprising a first generating means for generating the first word list.

8. A method for controlling an information processing device having a management unit for managing a format of a data item corresponding to a data type, comprising: a recognition execution step in which the recognition execution means acquires first text data that has been speech-recognized from the speech data using the first word list; a type identification step of identifying a data type of the voice data based on the acquired first text data by a type identification means; a value acquiring step in which a value acquiring means acquires a value of a data item corresponding to the specified data type from the first text data based on a format of the data item; Equipped with The recognition execution step includes obtaining second text data that is speech-recognized from the speech data using a second word list related to the specified data type; The value acquisition step acquires a value of a data item corresponding to the data type from the second text data based on a format of the data item.

23. A method for controlling an information processing apparatus comprising:

9. An information processing device, A management means for managing a format of a data item corresponding to a data type; a recognition execution means for acquiring first text data that is speech-recognized from the speech data using the first word list; a type identification means for identifying a data type of the voice data based on the acquired first text data; a value acquiring means for acquiring a value of a data item corresponding to the specified data type from the first text data based on a format of the data item; A program for causing the device to function as a the recognition execution means acquires second text data obtained by speech recognition from the speech data using a second word list related to the specified data type; The value acquiring means acquires a value of a data item corresponding to the data type from the second text data based on a format of the data item. A program characterized by:

Citation Information

Patent Citations

  • Word string recognizing device

    JP2001034292A

  • Method and device for speech recognition, speech recognition program, and storage medium with speech recognition program stored therein

    JP2005250071A

  • Speech recognition device

    JP2006039382A

  • Interactive device, method and program

    JP2016061954A

  • Dictionary update device and program

    JP2018040906A