Document creation device, program, and document creation method

Abstract

To enable creation of a document based on an unregistered word even in the case an unregistered word is input by voice in a document creation technique for creating a document in accordance with a predetermined format based on input voice.SOLUTION: For each of labels included in a template registered in association with category information received from a communication terminal 2, a report sentence creation and recording apparatus 1 determines whether or not a word matching any of words registered in association with that label is present among words extracted from voice information received from a communication terminal 2, and if such a word is present, replaces the label with the matching word. On the other hand, if there is no such a word, a word similar to any of words registered as replacement candidates for the label is selected from the words extracted from the voice information, and the label is replaced with the similar word. Thus, a report sentence is created.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a document creation technique for creating documents based on input speech, and more particularly to a document creation technique suitable for creating reports in accordance with a predetermined format. [Background technology]

[0002] Patent Document 1 discloses a template corpus creation device that creates a template corpus to be stored in a speech database and used to generate synthetic speech. This template corpus creation device includes a template input unit that accepts template words and the positions of arbitrary words to be input between the template words, a random word input unit that accepts the attributes of the arbitrary words to be inserted and their insertion positions, a random word selection unit that extracts words having the same attributes as the arbitrary words accepted by the random word input unit from among words registered in a word dictionary, a template generation unit that generates template sentences by inserting the words extracted by the random word selection unit between the template words accepted by the template input unit, and a template output unit that outputs and stores the template sentences generated by the template generation unit in the speech database as a template corpus. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2000-56787 Summary of the Invention [Problem to be solved by the invention]

[0004] In nursing care settings, in order to reduce the burden on caregivers, it is desirable to be able to create reports in a specified format based on voice input and record these reports in a nursing record database.

[0005] However, the fixed phrase corpus creation device described in Patent Document 1 is not intended to create documents based on voice input. Even if the fixed phrase input unit and the free word input unit in this fixed phrase corpus creation device were modified to accept voice input, the following problem would arise. That is, this fixed phrase corpus creation device generates fixed phrases by extracting words with the same attributes as the free words accepted by the free word input unit from among the words registered in the word dictionary and inserting them between the fixed words accepted by the fixed phrase input unit, so it is not possible to create fixed phrases using words not registered in the word dictionary.

[0006] The present invention has been made in consideration of the above circumstances, and its purpose is to enable a document creation technique for creating a document in a predetermined format based on input speech to create a document based on an unregistered word even when the unregistered word is input by speech. [Means for solving the problem]

[0007] To solve the above problem, the present invention pre-registers multiple templates containing labels to be replaced with words, each associated with category information. For each of the multiple templates, at least one word is registered as a replacement candidate for each label included in the template. When category information and speech information representing the report content are received, words of a predetermined part of speech are extracted from the speech information, a template associated with the received category information is selected, and the label included in the selected template is replaced with the word extracted from the speech information to create a document. For each label, the system determines whether any of the words registered as replacement candidates for that label are included in the words extracted from the speech information. If such a word is included, the label is replaced with the matching word. If such a word is not included in the words extracted from the speech information, a word similar to any of the words registered as replacement candidates for that label is selected from the words extracted from the speech information, and the label is replaced with the similar word.

[0008] For example, the present invention provides A document creation device that creates a document based on input voice, a template storage means for storing a plurality of templates each including a label to be replaced with a word in a sentence, the templates being associated with category information; a word dictionary means in which, for each of the plurality of templates, at least one word that is a candidate for replacing the label included in the template is registered; a category information receiving means for receiving the category information; a voice information receiving means for receiving voice information representing the report content; a word extraction means for extracting words of a predetermined part of speech from the voice information received by the voice information receiving means; a template selection means for selecting the template associated with the category information received by the category information reception means and stored in the template storage means; a document creation means for creating a document by replacing the label included in the template selected by the template selection means with the word extracted by the word extraction means, The document creation means For each label included in the template selected by the template selection means, it is determined whether or not a word matching any of the words registered in the word dictionary means as a replacement candidate for that label is included among the words extracted by the word extraction means, and if so, the label is replaced with the matching word, and if not, a word similar to any of the words registered in the word dictionary means as a replacement candidate for that label is selected from the words extracted by the word extraction means, and the label is replaced with the similar word. [Effects of the Invention]

[0009] In the present invention, for a label included in a template registered in association with the received category information, if there is a word extracted from the received speech information that matches one of the words registered in association with this label, the label is replaced with this matching word. On the other hand, if there is no such word, a word similar to one of the words registered as a replacement candidate for this label is selected from the words extracted from the speech information, and the label is replaced with this similar word.

[0010] Therefore, according to the present invention, even when an unregistered word is input by voice, a document in accordance with a predetermined format can be created based on this unregistered word. [Brief explanation of the drawings]

[0011] [Figure 1]FIG. 1 is a schematic diagram of a report writing and recording system according to an embodiment of the present invention. [Figure 2] FIG. 2 is a diagram showing a schematic functional configuration of the report writing and recording device 1. As shown in FIG. [Figure 3] FIG. 3 is a diagram showing an example of registered contents in the template storage unit 101. As shown in FIG. [Figure 4] FIG. 4 is a diagram showing an example of registered contents in the word dictionary unit 102. As shown in FIG. [Figure 5] FIG. 5 is a diagram showing an example of registered contents in the report storage unit 103. As shown in FIG. [Figure 6] FIG. 6 is a flow diagram for explaining the operation of the report writing and recording device 1. [Figure 7] FIG. 7 is a flow diagram for explaining the dictionary registration word replacement process S13 shown in FIG. [Figure 8] FIG. 8 is a flow diagram for explaining the dictionary registration word correction / replacement process S15 shown in FIG. [Figure 9] FIG. 9 is a flow diagram for explaining the dictionary unregistered word replacement process S17 shown in FIG. [Figure 10] FIG. 10 is a flow diagram for explaining the dictionary unregistered word replacement process S17 shown in FIG. 6, and is a continuation of FIG. DETAILED DESCRIPTION OF THE INVENTION

[0012] An embodiment of the present invention will be described below.

[0013] FIG. 1 is a schematic diagram of a report writing and recording system according to this embodiment.

[0014] As shown in the figure, the report writing and recording system according to this embodiment is used in nursing care settings and the like, and is configured by connecting a report writing and recording device 1 and at least one communication terminal 2 to a network 3. Note that while Fig. 1 illustrates a wired network such as a LAN (Local Area Network) or a WAN (Wide Area Network) as the network 3, the network 3 may also be a wireless network such as a wireless LAN or a mobile phone network.

[0015] The communication terminal 2 has a voice input / output function, and transmits category information received from a reporter such as a caregiver and voice information of the report content input via the network 3 to the report creation / recording device 1, and also receives and displays a report including the report from the report creation / recording device 1. Then, if necessary, the communication terminal 2 receives corrections to the report from the reporter and transmits the corrections to the report creation / recording device 1.

[0016] The communication terminal 2 may be a dedicated communication terminal such as an interphone, or may be a general-purpose communication terminal such as a mobile phone, a smartphone, or a tablet PC (Personal Computer) with a communication function.

[0017] The report creation and recording device 1 creates and records a report in accordance with a predetermined format (template) based on category information and voice information representing the report content received from the communication terminal 2 via the network 3. It also transmits the created report to the communication terminal 2 and accepts corrections to the report from the communication terminal 2.

[0018] FIG. 2 is a diagram showing a schematic functional configuration of the report writing and recording device 1. As shown in FIG.

[0019] As shown in the figure, the report creation and recording device 1 includes a network interface unit 100, a template storage unit 101, a word dictionary unit 102, a report storage unit 103, a category information receiving unit 104, a voice information receiving unit 105, a word extraction unit 106, a template selection unit 107, a report creation unit 108, a similarity measurement unit 109, a correction receiving unit 110, and a dictionary update unit 111.

[0020] The network interface unit 100 is an interface for connecting to the network 3 .

[0021] The template storage unit 101 stores templates that serve as the basis for report sentences in accordance with a predetermined format, linked to category information and verbs (in this case, the base form) used in the sentences.

[0022] FIG. 3 is a diagram showing an example of registered contents in the template storage unit 101. As shown in FIG.

[0023] As shown in the figure, template record 1010 is stored for each template in template storage unit 101. Template record 1010 has field 1011 in which a template ID for identifying the template is registered, field 1012 in which category information related to this template is registered, field 1013 in which verbs used in sentences included in this template are registered, and field 1014 in which this template is registered.

[0024] The template contains a sentence with one or more labels (strings of characters enclosed in <> in Figure 3) 1015 associated with the attributes of a word, and a report is created by replacing the label 1015 in this sentence with the word of the attribute associated with it.

[0025] The word dictionary unit 102 stores, for each template, words that are candidates for replacing each of the labels 1015 included in the template.

[0026] FIG. 4 is a diagram showing an example of registered contents in the word dictionary unit 102. As shown in FIG.

[0027] As shown in the figure, the word dictionary unit 102 stores, for each template stored in the template storage unit 101, a word dictionary table 1020 linked to the template ID of the template.

[0028] In the word dictionary table 1020, records 1021 of replacement candidates are registered for each label 1015 included in a template identified by a template ID linked to this table 1020.

[0029] The replacement candidate record 1021 has a field 1022 in which the attribute of the label 1015 is registered, and a field 1023 in which at least one word that is a replacement candidate for this label 1015 is registered.

[0030] The report storage unit 103 stores reports created by the report creation unit 108.

[0031] FIG. 5 is a diagram showing an example of registered contents in the report storage unit 103. As shown in FIG.

[0032] As shown in the figure, a report record 1030 is stored for each report in the report storage unit 103. The report record 1030 has a field 1031 in which the report date and time is registered, a field 1032 in which a terminal ID that is identification information of the communication terminal 2 of the reporter (e.g., a caregiver) is registered, and a field 1033 in which a report for an object (e.g., a care recipient) that the reporter is responsible for is registered.

[0033] The category information receiving unit 104 receives category information from the communication terminal 2 via the network interface unit 100. The category information receiving unit 104 may receive voice information of the category information received from the communication terminal 2 and acquire the category information by performing voice recognition on the voice information.

[0034] The voice information receiving unit 105 receives voice information representing the report content from the communication terminal 2 via the network interface unit 100.

[0035] The word extraction unit 106 extracts verbs and nouns contained in the report content represented by the voice information received by the voice information receiving unit 105. Specifically, the word extraction unit 106 performs a voice recognition process on the voice information to convert it into text data, and then performs a morphological analysis process on the text data to break it down into multiple words. Then, the word extraction unit 106 extracts verbs and nouns from the multiple words obtained by breaking down the text data.

[0036] The template selection unit 107 selects a template stored in the template storage unit 101 in association with the category information received by the category information reception unit 104 and the verb extracted by the word extraction unit 106 .

[0037] The report creation unit 108 works in cooperation with the similarity measurement unit 109 to create a report using the template selected by the template selection unit 107, a word dictionary table 1020 linked to the template ID of this template and stored in the word dictionary unit 102, and nouns extracted by the word extraction unit 106.

[0038] The similarity measurement unit 109 measures the similarity between words in accordance with instructions from the report creation unit 108. For example, it measures the similarity between words input from the report creation unit 108 using a trained model such as Word2Vec that has been trained using general documents. Here, the trained model does not necessarily have to be included in the report creation and recording device 1 itself, and a trained model included in another AI (Artificial Intelligence) server connected to the network 3 may be used.

[0039] Furthermore, the similarity measurement unit 109 measures the pronunciation similarity between words in accordance with instructions from the report creation unit 108. For example, it measures the Levenshtein distance between words input from the report creation unit 108 (the smaller the distance, the higher the pronunciation similarity between words).

[0040] The correction receiving unit 110 receives an instruction to correct the report created by the report creating unit 108 from the communication terminal 2 that is the sender of the voice information used to create this report.

[0041] The dictionary update unit 111, in accordance with instructions from the report creation unit 108, registers unregistered words (nouns) used in creating the report in a table 1020 of the word dictionary used in creating the report.

[0042] The functional configuration of the report creation and recording device 1 shown in Fig. 2 may be realized in hardware using an integrated logic IC such as an ASIC (Application Specific Integrated Circuit) or FPGA (Field Programmable Gate Array), or in software using a computer such as a DSP (Digital Signal Processor). Alternatively, it may be realized as a process in a computer system such as a PC equipped with a CPU, memory, an auxiliary storage device such as a HDD or DVD-ROM, and a communication interface such as a modem, NIC, or wireless LAN adapter, by the CPU loading a predetermined program from the auxiliary storage device into memory and executing it.

[0043] FIG. 6 is a flow diagram for explaining the operation of the report writing and recording device 1.

[0044] This flow starts when the category information receiving unit 104 receives category information and the voice information receiving unit 105 receives voice information representing the report content from the same communication terminal 2 via the network interface unit 100 .

[0045] First, the category information receiving unit 104 links the category information to the communication terminal 2 that is the sender of the category information and outputs the linked category information to the template selecting unit 107. Furthermore, the voice information receiving unit 105 links the voice information to the communication terminal 2 that is the sender of the voice information and outputs the voice information to the word extracting unit 106.

[0046] Next, the word extraction unit 106 performs speech recognition processing on the speech information received by the speech information receiving unit 105 to convert the speech information into text data, and then performs morphological analysis processing on this text data to break it down into multiple words.The word extraction unit 106 then extracts verbs and nouns from these words, links the extracted verbs (hereinafter referred to as extracted verbs) to the communication terminal 2 linked to the speech information and outputs them to the template selection unit 107, and links the extracted nouns (hereinafter referred to as extracted nouns) to the communication terminal 2 and outputs them to the report creation unit 108 (S10).

[0047] Next, when template selection unit 107 receives category information and extracted verbs linked to the same communication terminal 2 from category information reception unit 104 and word extraction unit 106, respectively, template selection unit 107 selects a template linked to the category information and extracted verb and stored in template storage unit 101. Then, the selected template (hereinafter referred to as the selected template) and its template ID are output to report creation unit 108, linked to this communication terminal 2 (S11).

[0048] Next, when the report creation unit 108 receives extracted nouns and selected templates, etc. linked to the same communication terminal 2 from the word extraction unit 106 and the template selection unit 107, respectively, it selects a word dictionary table (hereinafter referred to as the selected dictionary table) 1020 linked to the template ID of the selected template and stored in the word dictionary unit 102 (S12).

[0049] Thereafter, the report creation unit 108 performs a dictionary registration word replacement process S13, which will be described later, to replace the label 1015 included in the selected template with an extracted noun that is registered as a replacement candidate for this label in the selected dictionary table 1020. As a result, if all the labels 1015 included in the selected template have been replaced with extracted nouns (YES in S14), the process proceeds to S19, and if there is an unreplaced label 1015 that has not been replaced with an extracted noun (NO in S14), the report creation unit 108 performs a dictionary registration word correction / replacement process S15, which will be described later, to correct the extracted noun whose pronunciation matches or is similar to the replacement candidate for the unreplaced label registered in the selected dictionary table 1020 to this replacement candidate, and replace this unreplaced label with the corrected extracted noun. As a result, if all the labels 1015 included in the selected template have been replaced with extracted nouns (YES in S16), the process proceeds to S19. If there are any unreplaced labels that have not been replaced with extracted nouns (NO in S16), the unregistered word replacement process S17 described below is performed to replace the unreplaced labels with extracted nouns that are similar to replacement candidates for these unreplaced labels registered in the selected dictionary table 1020.

[0050] Next, the report creation unit 108 determines whether the created report contains any unsubstituted labels that have not been replaced with extracted nouns (S18). If all of the labels 1015 included in the selected template have been replaced with extracted nouns (NO in S18), the process proceeds to S19. If any unsubstituted labels are included (YES in S18), the process performs a predetermined error process, such as sending a message via the network interface unit 100 to the communication terminal 2 linked to the extracted noun and the selected template, stating that the report creation has failed and prompting the user to re-enter category information and audio information about the report content (S25).

[0051] When all labels 1015 included in the selected template have been replaced as described above and the creation of the report for the operator (reporter) of communication terminal 2 that is the sender of the category information and voice information is complete (YES in S14, YES in S16, or NO in S18), in S19, report creation unit 108 instructs correction receiving unit 110 to transmit the report to communication terminal 2 that is the sender of the category information and voice information. Correction receiving unit 110 transmits the report created by report creation unit 108 to communication terminal 2 that is the sender of the category information and voice information via network interface unit 100, causes the report to be displayed, and prompts the reporter to check the report. Then, a recording instruction from the reporter is received from this communication terminal 2.

[0052] Next, if the content of the report correction is added to the recording instruction received by the correction receiving unit 110 from the communication terminal 2 (YES in S20), the report creating unit 108 corrects the report in accordance with the content of the correction. Then, the report including the corrected report is recorded in the report storage unit 103 in association with the report date and time and the terminal ID of the communication terminal 2 that sent the recording instruction (S21).

[0053] On the other hand, if the recording instruction received by the correction receiving unit 110 from the communication terminal 2 does not include any corrections to the report (NO in S20), the report creation unit 108 records the report including the report in the report storage unit 103, linking it to the report date and time and the terminal ID of the communication terminal 2 that sent the recording instruction (S22). Then, if the extracted noun that replaced the label 1015 included in the selected template in creating the report contains a word that is not registered in the selected dictionary table 1020 (YES in S23), the dictionary update unit 111 adds this unregistered word to the selected dictionary table 1020 as a replacement candidate for this label, and updates the word dictionary unit 102 (S24).

[0054] FIG. 7 is a flow diagram for explaining the dictionary registration word replacement process S13 shown in FIG.

[0055] First, the report creation unit 108 selects an unselected label 1015 from the labels 1015 included in the selected template (S130), and identifies a replacement candidate linked to this label (hereinafter referred to as the selected label in the dictionary registration word replacement process S13) 1015 from the selected dictionary table 1020 (S131).

[0056] Next, the report creation unit 108 requests the similarity measurement unit 109 to measure the similarity between all the identified replacement candidates and all the extracted nouns that do not have a replaced flag (described later) attached to them, by passing them to the similarity measurement unit 109. In response to this, the similarity measurement unit 109 measures the similarity between each replacement candidate and each extracted noun using, for example, a trained model trained from general documents, and passes the measurement results to the report creation unit 108 (S132).

[0057] The report creation unit 108 then refers to the measurement results received from the similarity measurement unit 109 and determines whether or not there is an extracted noun that matches any of the replacement candidates (S133). If there is an extracted noun that matches any of the replacement candidates (YES in S133), the report creation unit 108 replaces the selected label 1015 in the selection template with this extracted noun (S134), and assigns a replaced flag to this extracted noun (S135).

[0058] Next, if there is an unselected label 1015 in the selected template (YES in S136), the report creation unit 108 returns to S130 and continues processing; on the other hand, if there is no unselected label 1015 in the selected template (NO in S136), the report creation unit 108 ends this flow and proceeds to S14 in Figure 6.

[0059] FIG. 8 is a flow diagram for explaining the dictionary registration word correction / replacement process S15 shown in FIG.

[0060] First, the report creation unit 108 selects an unselected label 1015 from among the unreplaced labels 1015 remaining in the selected template after the dictionary registration word replacement process S13 is performed (S150), and identifies a replacement candidate linked to this label (hereinafter referred to as the selected label in the dictionary registration word correction / replacement process S15) 1015 from the selected dictionary table 1020 (S151).

[0061] Next, the report creation unit 108 requests the similarity measurement unit 109 to measure the pronunciation similarity between all the identified replacement candidates and all the extracted nouns that have not been flagged as replaced, and passes the measurement results to the report creation unit 108 (S152).

[0062] Next, the report creation unit 108 refers to the measurement result received from the similarity measurement unit 109 and determines whether there is an extracted noun whose pronunciation matches any of the replacement candidates (S153). If there is an extracted noun whose pronunciation matches any of the replacement candidates (YES in S153), the report creation unit 108 corrects the extracted noun to a replacement candidate whose pronunciation matches that of the extracted noun (S154). On the other hand, if there is an extracted noun whose pronunciation does not match any of the replacement candidates (NO in S153), the report creation unit 108 further determines whether there is an extracted noun among those extracted nouns whose pronunciation similarity with any of the replacement candidates is higher than a predetermined level (specifically, an extracted noun whose Levenshtein distance is equal to or less than a predetermined value) (S155). If there is an extracted noun whose pronunciation similarity with any of the replacement candidates is higher than a predetermined level (YES in S155), the extracted noun is corrected to a replacement candidate whose pronunciation similarity with the extracted noun is higher than a predetermined level (S156).

[0063] Then, the report creation unit 108 replaces the selected label 1015 in the selected template with the corrected extracted noun (S157), and assigns a replacement flag to this corrected extracted noun (S158).

[0064] Thereafter, if there is an unselected label 1015 in the selected template (YES in S159), the report creation unit 108 returns to S150 and continues processing; on the other hand, if there is no unselected label 1015 in the selected template (NO in S159), the report creation unit 108 ends this flow and proceeds to S16 in Figure 6.

[0065] 9 and 10 are flow diagrams for explaining the dictionary unregistered word replacement process S17 shown in FIG.

[0066] First, the report creation unit 108 compares the number of extracted nouns to which the replacement flag has not been assigned with the number of unreplaced labels 1015 that have not been replaced with extracted nouns (S170). If the number of extracted nouns to which the replacement flag has not been assigned is smaller than the number of unreplaced labels 1015 (NO in S170), the process proceeds to S18 in FIG.

[0067] On the other hand, if the number of extracted nouns that have not been assigned a replacement flag is greater than or equal to the number of unreplaced labels 1015 (YES in S170), the report creation unit 108 identifies, for each unreplaced label 1015, a replacement candidate word linked to the unreplaced label 1015 from the selected dictionary table 1020 (S171).

[0068] Next, the report creation unit 108 creates replacement patterns for all combinations of unreplaced labels 1015 and extracted nouns that have not been assigned a replacement flag, in which any extracted noun that has not been assigned a replacement flag is assigned to each unreplaced label 1015 in the selected template, so that duplicate extracted nouns are not used between labels (S172).

[0069] The report creation unit 108 then selects an unselected replacement pattern from the created replacement patterns (S173). Then, for each unsubstituted label 1015 included in the selected replacement pattern, the report creation unit 108 passes the extracted noun assigned to this label 1015 and each of the replacement candidates for this label 1015 to the similarity measurement unit 109, requesting it to measure the similarity between them. In response to this, the similarity measurement unit 109 measures the similarity between the extracted noun assigned to this label 1015 and each of the replacement candidates for this label 1015 for each unsubstituted label 1015 using the trained model, and passes the measurement results to the report creation unit 108 (S174).

[0070] Next, the report creation unit 108 refers to the measurement result received from the similarity measurement unit 109 and determines whether or not there is a label 1015 among the unsubstituted labels 1015 included in the selected replacement pattern for which the maximum similarity between the extracted noun assigned to this label 1015 and each of the replacement candidates for this label 1015 is equal to or less than a predetermined value (S175). If such an unsubstituted label 1015 exists (YES in S175), the report creation unit 108 sets the similarity of the selected replacement pattern to "zero" (S176). On the other hand, if such an unsubstituted label 1015 does not exist (NO in S175), for each unsubstituted label 1015 included in the selected replacement pattern, the report creation unit 108 identifies the maximum similarity between the extracted noun assigned to this label 1015 and each of the replacement candidates for this label 1015, and sets the sum of these similarities as the similarity of the selected replacement pattern (S177).

[0071] Next, if there is an unselected replacement pattern among the replacement patterns that have been created (NO in S178), the report creation unit 108 returns to S173 and continues the process.

[0072] On the other hand, if all the created replacement patterns have been selected (YES in S178), it is determined whether or not any of these replacement patterns has a similarity equal to or greater than a predetermined reference value (S179). If no replacement pattern has a similarity equal to or greater than the predetermined reference value (NO in S179), it is determined that there is no extracted noun that can be substituted for the unsubstituted label 1015 among the extracted nouns to which the replaced flag has not been assigned, and the process proceeds to S18 in FIG.

[0073] On the other hand, if there are replacement patterns whose similarity is equal to or greater than the predetermined reference value (YES in S179), the report creation unit 108 selects the replacement pattern with the greatest similarity from among such replacement patterns as the adopted pattern (S180). Then, according to this adopted pattern, it replaces each unsubstituted label 1015 included in the selected template with the extracted noun assigned to this label 1015 (S181), and assigns a replaced flag to this extracted noun (S182).

[0074] For example, suppose the selected template is "I ate <amount> of <food> at <amount>" which includes labels 1015 of the attributes <amount>, <location>, and <amount>, and the replacement candidates for the label 1015 of the attribute <amount> registered in the selected dictionary table 1020 are "private room", "room", and "dining room", the replacement candidates for the label 1015 of the attribute <food> registered in the selected dictionary table 1020 are "breakfast", "lunch", "dinner", and "snack", and the replacement candidates for the label 1015 of the attribute <amount> registered in the selected dictionary table 1020 are "full amount", "half amount", and "small amount".

[0075] Assume that the extracted nouns "private room," "breakfast," and "full amount" are extracted from the speech information. In this case, the replacement candidate "private room" for the label 1015 of the attribute <location> matches the extracted noun "private room," the replacement candidate "breakfast" for the label 1015 of the attribute <food> matches the extracted noun "breakfast," and the replacement candidate "full amount" for the label 1015 of the attribute <quantity> matches the extracted noun "full amount." Therefore, by the dictionary registration word replacement process S13 shown in FIG. 7, the labels 1015 of the attributes <location>, <food>, and <quantity> included in the selected template are replaced with the extracted nouns "private room," "breakfast," and "full amount," respectively. As a result, the report sentence "I ate the full amount of breakfast in a private room" is created.

[0076] Also, suppose the extracted nouns "dining hall," "champion," and "goodness" are extracted from the speech information. In this case, the replacement candidate "dining hall" for label 1015 of the attribute <location> matches the extracted noun "dining hall." Therefore, the label 1015 of the attribute <location> included in the selected template is replaced with the extracted noun "dining hall" by the dictionary registration word replacement process S13 shown in FIG. 7. Furthermore, none of the replacement candidates for label 1015 of the attribute <food> match the extracted nouns "champion" or "goodness," but the replacement candidate "dining hall" differs in pronunciation from the extracted noun "champion" by only one character, so the two are similar in pronunciation. Therefore, the extracted noun "dining hall" is corrected to "dining hall" by the dictionary registration word correction / replacement process S15 shown in FIG. 8, and the unreplaced label 1015 of the attribute <food> included in the selected template is replaced with the corrected extracted noun "dining hall." Furthermore, none of the replacement candidates for the attribute <quantity> label 1015 match the extracted noun "good", but the replacement candidate "full amount" matches the extracted noun "good". Therefore, the dictionary registration word correction / replacement process S15 shown in Figure 8 corrects the extracted noun "good" to "full amount", and the unreplaced label 1015 for the attribute <quantity> included in the selected template is replaced with the corrected extracted noun "full amount". As a result, the report "I ate the full amount of dinner at the cafeteria." is created.

[0077] Also, suppose that the extracted nouns "dining," "lunch," and "full amount" are extracted from the speech information. In this case, the replacement candidate "full amount" for the label 1015 of the attribute <quantity> matches the extracted noun "full amount." Therefore, the label 1015 of the attribute <quantity> included in the selected template is replaced with the extracted noun "full amount" by the dictionary registration word replacement process S13 shown in FIG. 7. Furthermore, none of the replacement candidates for the labels 1015 of the attributes <location> and <food> match the extracted nouns "dining" and "lunch," and their pronunciations are also not similar. Therefore, the dictionary non-registered word replacement process S17 shown in FIGS. 9 and 10 is performed, and the similarity is calculated between a replacement pattern in which the extracted nouns "dining" and "lunch" are assigned to the unsubstituted labels 1015 of the attributes <location> and <food> in the selected template, respectively, and a replacement pattern in which the extracted nouns "lunch" and "dining" are assigned to them, respectively. Here, the extracted noun "dining" has a high similarity to the replacement candidate "dining hall" of the label 1015 of the attribute <location>, and the extracted noun "lunch" has a high similarity to the replacement candidate "lunch" of the label 1015 of the attribute <food>. Therefore, the replacement pattern in which the extracted nouns "dining" and "lunch" are assigned to the labels 1015 of the attributes <location> and <food>, respectively, is determined to be the adopted pattern, and the unsubstituted labels 1015 of the attributes <location> and <food> included in the selected template are replaced with the extracted nouns "dining" and "lunch", respectively. As a result, the report "I ate all of my lunch at the dining room." is created.

[0078] One embodiment of the present invention has been described above.

[0079] In this embodiment, if the extracted nouns extracted from the speech information contain an extracted noun that matches a replacement candidate for a label included in a selected template, the report creation and recording device 1 replaces the label with the matching extracted noun. On the other hand, if there is no such extracted noun, and if the extracted nouns extracted from the speech information contain an extracted noun that is similar to one of the replacement candidates for this label, the report creation and recording device 1 replaces the label with the similar extracted noun. Therefore, according to this embodiment, even if an unregistered word is input by speech as a replacement candidate, a report can be created in accordance with a predetermined format based on this unregistered word.

[0080] Furthermore, in this embodiment, when the selected template contains one or more unsubstituted labels for which no extracted nouns match any replacement candidates, the report creation and recording device 1 selects an extracted noun for each unsubstituted label so that the combination of the unsubstituted label and the extracted noun maximizes the total maximum similarity (similarity of replacement patterns) between the extracted noun and the words registered as replacement candidates for the unsubstituted label, and replaces the unsubstituted label with the selected extracted noun. Therefore, according to this embodiment, even when the speech information contains one or more unregistered words as replacement candidates, a more accurate report can be created based on the unregistered words.

[0081] Furthermore, in this embodiment, the report creation and recording device 1 corrects extracted nouns whose pronunciation matches or is similar to a replacement candidate label in a selected template by a predetermined degree or more. Therefore, according to this embodiment, even if an extracted noun cannot be correctly extracted from speech information due to a recognition error in speech recognition processing, a more accurate report can be created.

[0082] Furthermore, in this embodiment, the report creation and recording device 1 stores templates linked to category information and verbs used in the sentence, and upon receiving category information and audio information of the report content from the communication terminal 2, determines the template linked to the category information and the extracted verb from the audio information as the selected template. Therefore, according to this embodiment, it is possible to appropriately select a template to serve as the basis for a report from one or more stored templates.

[0083] Furthermore, in this embodiment, when a label of a selected template is converted into an extracted noun that does not match any of the replacement candidates for this label, the report writing and recording device 1 adds and registers this extracted noun as a replacement candidate for the label of this selected template. Therefore, according to this embodiment, every time a word that is not registered as a replacement candidate is input by voice, this word is automatically added to the word dictionary as a replacement candidate, thereby improving convenience.

[0084] The present invention is not limited to the above-described embodiment, and various modifications are possible within the scope of the present invention.

[0085] For example, in the above embodiment, the report creation and recording device 1 stores templates linked to category information and verbs used in the sentences. Upon receiving category information and speech information of the report content from the communication terminal 2, the device extracts verbs from the speech information and determines the template linked to the category information and the extracted verb from the speech information as the selected template. Therefore, if a verb cannot be extracted from the speech information by speech recognition processing or morphological analysis processing, or if the speech information does not contain a verb, it may not be possible to determine a single selected template based on category information alone, and multiple selected templates may be determined. In such cases, a report may be created for each selected template, as described below, and the most appropriate report may be selected from the created reports. That is, when multiple templates are determined as selected templates, the report creation and recording device 1 creates a report for each selected template according to the above-described procedure. Then, for each report sentence created, the system measures the similarity between the extracted noun that replaced this label and the replacement candidate for this label for each label included in the selection template that forms the basis of the report sentence (if the extracted noun and the replacement candidate match, the similarity between them is set to the highest value), and calculates the sum of the maximum similarities measured for each label.Then, the report sentence with the highest total similarity value is sent to the communication terminal 2 as the optimal report sentence and displayed, and the reporter is allowed to check this report sentence.

[0086] Furthermore, in the above embodiment, the report creation and recording device 1 stores templates linked to category information and verbs used in the sentence, receives category information and audio information of the report content from the communication terminal 2, extracts verbs from the audio information, and determines the template linked to the category information and the extracted verb from the audio information as the selected template. However, the present invention is not limited to this. Templates may be stored linked to category information, and the category information may be received from the communication terminal 2, and the template linked to this category information may be determined as the selected template. In this case, the category information may be subdivided so that one template is linked to one piece of category information.

[0087] In the above embodiment, the report creation and recording device 1 extracts nouns from the speech information received from the communication terminal 2 and replaces each label included in the selected template with the extracted noun from the speech information. However, the present invention is not limited to this. The present invention may extract words of a predetermined part of speech from the speech information and replace each label included in the selected template with an extracted word from the speech information. For example, adjectives as well as nouns may be extracted from the speech information, and if there is an extracted word (here, an extracted noun or extracted adjective) that matches one of the label replacement candidates, the label is replaced with the matching extracted word. On the other hand, if there is no such extracted word, and there is an extracted word similar to one of the label replacement candidates or an extracted word with a similar pronunciation, the label is replaced with the extracted word. [Explanation of symbols]

[0088] 1: Report writing and recording device 2: Communication terminal 3: Network 100: Network interface unit 101: Template storage unit 102: Word dictionary section 103: Report text storage section 104: Category information reception unit 105: Voice information reception unit 106: Word extraction unit 107: Template selection unit 108: Report writing section 109: Similarity measurement section 110: Correction reception unit 111: Dictionary update unit

Claims

1. A document creation device that creates a document based on input voice, a template storage means for storing a plurality of templates each including a label to be replaced with a word in a sentence, the templates being associated with category information; a word dictionary means in which, for each of the plurality of templates, at least one word that is a candidate for replacing the label included in the template is registered; category information receiving means for receiving the category information; a voice information receiving means for receiving voice information representing the report content; a word extraction means for extracting words of a predetermined part of speech from the speech information received by the speech information receiving means; a template selection means for selecting the template associated with the category information received by the category information reception means and stored in the template storage means; a document creation means for creating a document by replacing the label included in the template selected by the template selection means with the word extracted by the word extraction means, The document creation means For each label included in the template selected by the template selection means, it is determined whether or not a word that matches any of the words registered in the word dictionary means as a replacement candidate for that label is included among the words extracted by the word extraction means, and if so, the label is replaced with the matching word, and if not, a word similar to any of the words registered in the word dictionary means as a replacement candidate for that label is selected from the words extracted by the word extraction means, and the label is replaced with the similar word. A document creation device characterized by:

2. 2. The document creation device according to claim 1, The document creation means If the words extracted by the word extraction means do not include a word that matches any of the words registered in the word dictionary means as replacement candidates for one or more labels included in the template selected by the template selection means, a word extracted by the word extraction means is selected for each of the one or more labels, and the one or more labels are replaced with the selected word, so that the combination of the one or more labels and the word extracted by the word extraction means maximizes the total maximum value of similarities between the words registered in the word dictionary means as replacement candidates for each of the one or more labels and the word extracted by the word extraction means. A document creation device characterized by:

3. 2. The document creation device according to claim 1, The document creation means If the words extracted by the word extraction means include a word whose pronunciation matches or is similar to at least a predetermined level to any of the words registered in the word dictionary means as replacement candidates for the label included in the template selected by the template selection means, the word extracted by the word extraction means is corrected to the word whose pronunciation matches or is similar to at least a predetermined level to the word registered in the word dictionary means as replacement candidates for the label included in the template selected by the template selection means. A document creation device characterized by:

4. 2. The document creation device according to claim 1, The document creation means When a plurality of templates are selected by the template selection means, for each of the selected plurality of templates, the label included in the template is replaced with the word extracted by the word extraction means to create a document, and for each label included in the template, the maximum value of the similarity between the word into which the label has been replaced and a word registered in the word dictionary means as a replacement candidate for the label is measured, and the document created using the template with the highest total value of the measured maximum similarity is output. A document creation device characterized by:

5. 5. The document creation device according to claim 1, The template storage means the template is stored in association with the category information and the verb used in the sentence, The word extraction means extracting verbs in addition to the words of the predetermined parts of speech from the speech information received by the speech information receiving means; The template selection means Select the template stored in the template storage means in association with the category information received by the category information receiving means and the verb extracted by the word extracting means. A document creation device characterized by:

6. 5. The document creation device according to claim 1, The system further comprises a dictionary update means for, when the label of the template selected by the template selection means is replaced by a word other than a word registered in the word dictionary means as a replacement candidate for the label by the document creation means, linking the replaced word to the label and registering it in the word dictionary means. A document creation device characterized by:

7. A program that causes a computer to function as a document creation device that creates a document based on input voice, a template storage means for storing a plurality of templates each including a label to be replaced with a word in a sentence, the templates being associated with category information; a word dictionary means in which, for each of the plurality of templates, at least one word that is a candidate for replacing the label included in the template is registered; a category information receiving means for receiving the category information; a voice information receiving means for receiving voice information representing the report content; a word extraction means for extracting words of a predetermined part of speech from the speech information received by the speech information receiving means; a template selection means for selecting the template stored in the template storage means in association with the category information received by the category information reception means; and causing the computer to function as document creation means for creating a document by replacing the label included in the template selected by the template selection means with the word extracted by the word extraction means; The document creation means For each label included in the template selected by the template selection means, it is determined whether or not a word that matches any of the words registered in the word dictionary means as a replacement candidate for that label is included among the words extracted by the word extraction means, and if so, the label is replaced with the matching word, and if not, a word similar to any of the words registered in the word dictionary means as a replacement candidate for that label is selected from the words extracted by the word extraction means, and the label is replaced with the similar word. A program characterized by:

8. A document creation method for creating a document based on input speech, comprising: receiving voice information representing the report content and extracting words of a predetermined part of speech from the voice information; Accepting category information, selecting a template that includes a label to be replaced with a word in a sentence, the label being associated with the category information and stored in advance, For the label included in the selected template, if a word that matches any of the words registered in advance as a replacement candidate for the label is included in the words extracted from the speech information, the label is replaced with the matching word; if not, a word similar to any of the words registered in advance as a replacement candidate for the label is selected from the extracted words, and the label is replaced with the similar word, thereby creating a document. A document creation method comprising:

Images