WO2017020657A1 - 从语音信息中识别时间信息的方法和装置 - Google Patents

从语音信息中识别时间信息的方法和装置 Download PDF

Info

Publication number
WO2017020657A1
WO2017020657A1 PCT/CN2016/086125 CN2016086125W WO2017020657A1 WO 2017020657 A1 WO2017020657 A1 WO 2017020657A1 CN 2016086125 W CN2016086125 W CN 2016086125W WO 2017020657 A1 WO2017020657 A1 WO 2017020657A1
Authority
WO
WIPO (PCT)
Prior art keywords
time
feature
information
field
intermediate data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2016/086125
Other languages
English (en)
French (fr)
Inventor
王飞
徐浩
冯晓冰
许振兴
裴国东
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tencent Technology Shenzhen Co Ltd
Original Assignee
Tencent Technology Shenzhen Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tencent Technology Shenzhen Co Ltd filed Critical Tencent Technology Shenzhen Co Ltd
Priority to EP16832145.3A priority Critical patent/EP3330962B1/en
Priority to KR1020187000743A priority patent/KR102046728B1/ko
Priority to JP2018504856A priority patent/JP6667931B2/ja
Publication of WO2017020657A1 publication Critical patent/WO2017020657A1/zh
Priority to US15/698,423 priority patent/US10249302B2/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/30—Semantic analysis
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/26—Speech to text systems
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
    • G10L2015/223—Execution procedure of a spoken command
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
    • G10L2015/226—Procedures used during a speech recognition process, e.g. man-machine dialogue using non-speech characteristics
    • G10L2015/228—Procedures used during a speech recognition process, e.g. man-machine dialogue using non-speech characteristics of application context

Definitions

  • the present application relates to the field of speech recognition technology, and in particular, to a method and apparatus for recognizing time information from voice information.
  • speech recognition is incorporated into the terminal application, and search queries or operational functions are implemented through voice interaction in the terminal, such as querying weather, flight dynamics, searching for surrounding food, making calls, setting reminders, adjusting alarm clocks. , playing music, etc.
  • voice interaction such as querying weather, flight dynamics, searching for surrounding food, making calls, setting reminders, adjusting alarm clocks. , playing music, etc.
  • the voice command often carries time information, such as reminding the meeting at 10 am on May 5, inquiring about the flight from Shenzhen to Beijing next Monday, and setting an alarm clock at 8 o'clock tomorrow morning.
  • the embodiment of the present invention provides a method and a device for identifying time information from voice information.
  • the program code is not required to be modified or recompiled, and the operability is strong.
  • the embodiment of the present application provides a method for identifying time information from voice information, including the following steps:
  • the intermediate data includes a plurality of fields, each field corresponding to one of the configuration files;
  • Time data corresponding to the time information in the voice information is obtained according to content of each field in the intermediate data.
  • An embodiment of the present application further provides an apparatus for identifying time information from voice information, including:
  • a voice collection module configured to collect voice information, and identify voice information including time information to obtain text information corresponding to the voice information
  • Configuring a query module configured to identify at least one time feature that matches the text information, and search for at least one configuration file according to the matched time feature to obtain a corresponding time character
  • An intermediate value-adding module configured to assign the corresponding time character to a field in the intermediate data corresponding to the configuration file in which the matched time feature is located; the intermediate data includes a plurality of fields, each field corresponding to one of the Configuration file
  • a time conversion module configured to obtain time data corresponding to the time information in the voice information according to content of each field in the intermediate data.
  • the time feature that may appear in the voice information is put into the configuration file of the corresponding category, and the configuration file of each category is corresponding to each field in the intermediate data, and when the time information in the voice information is recognized, Find matching time characteristics in the configuration file, and Obtaining a time character corresponding to the matching time feature, assigning the corresponding time character to the corresponding field in the intermediate data, and generating exact time data according to the content of the field, so that all key time features to be matched are stored in
  • the recognition program only needs to extract the time character corresponding to the time feature from the configuration file, and obtain the exact time data through the intermediate file conversion, so that the program code is more readable and effectively avoids the length of the program code.
  • the time feature needs to be modified or added, only the configuration file needs to be modified, and the program code is not modified and re-encoded, which simplifies the modification process of the identification program and provides convenience for the maintenance of the program code.
  • 1 is a bus diagram of a terminal where a device for identifying time information from voice information is located in an embodiment of the present application;
  • FIG. 2 is a flowchart of a first embodiment of a method for identifying time information from voice information according to the present application
  • FIG. 3 is a flowchart of a second embodiment of a method for identifying time information from voice information according to the present application
  • FIG. 4 is a flowchart of a third embodiment of a method for identifying time information from voice information according to the present application
  • FIG. 5 is a flowchart of a fourth embodiment of a method for identifying time information from voice information according to the present application
  • FIG. 6 is a flowchart of a fifth embodiment of a method for identifying time information from voice information according to the present application.
  • FIG. 7 is a flowchart of a sixth embodiment of a method for identifying time information from voice information according to the present application.
  • FIG. 8 is a schematic diagram of a first embodiment of an apparatus for identifying time information from voice information according to the present application.
  • FIG. 9 is a schematic block diagram of a second embodiment of an apparatus for identifying time information from voice information according to the present application.
  • FIG. 10 is a schematic diagram of a first embodiment of a configuration file in the present application.
  • Figure 11 is a schematic diagram of a second embodiment of a configuration file in the present application.
  • the hard information is used to identify the time information in the voice command. Specifically, various keywords are encoded in the program code, and after the keywords in the voice command are recognized, they are converted into corresponding time parameters to generate corresponding time.
  • this hard coding method requires all time keywords to be added to the code.
  • FIG. 1 is a bus diagram of a terminal where a device for identifying time information is received from voice information according to an embodiment of the present application.
  • the terminal may be a mobile phone, a pad, a computer, etc., and may include at least one processor 101, such as a CPU, at least one network interface 104, a user interface 103, a memory 105, and at least one communication bus 102.
  • the communication bus 102 is used to implement connection communication between these components.
  • the user interface 103 may include a display, a keyboard, and may also include a standard wired interface and a wireless interface.
  • Network interface 104 may include a standard wired interface, a wireless interface (such as a WIFI interface).
  • the memory 105 may be a high speed RAM memory or a non-volatile memory such as at least one disk memory.
  • the memory 105 can also be at least one storage device located remotely from the aforementioned processor 101.
  • Memory as a computer storage medium
  • the operating system, the network communication module, the user interface module, and the time information program from the voice information may be included in the 105.
  • the voice information may be natural voice information, such as voice information recognized from a language expressed by a person.
  • the network interface 104 is mainly used to connect to a server or other terminal, and performs data communication with a server or other terminal; and the user interface 103 is mainly used for receiving user commands. And interacting with the user; and the processor 101 can be used to call the program for identifying the time information from the voice information stored in the memory 105, and perform the following operations:
  • the intermediate data includes a plurality of fields, each field corresponding to a configuration file;
  • Time data corresponding to time information in the voice information is obtained according to the content of each field in the intermediate data.
  • each field in the intermediate data may be: a field corresponding to each configuration file.
  • the processor 101 invokes the program for identifying time information from the voice information stored in the memory 105.
  • the program may also perform the following operations:
  • the time character corresponding to the new time feature is set according to the time character of the same time feature
  • the processor 101 invokes the program for identifying time information from the voice information stored in the memory 105.
  • the program may also perform the following operations:
  • the time character corresponding to the matched time feature is found in the time-staff table in the configuration file in which the matching time feature is located.
  • the processor 101 invokes the program for identifying time information from the voice information stored in the memory 105.
  • the program may also perform the following operations:
  • the current time is obtained when the AM/PM field is not assigned in the intermediate data that has been assigned, and the date field is not assigned or the same date as the current time;
  • the AM/PM field of the intermediate data is set to coincide with the current time period
  • the AM/PM field of the intermediate data is set to be opposite to the current time period.
  • the processor 101 invokes the program for identifying time information from the voice information stored in the memory 105.
  • the program may also perform the following operations:
  • the processor 101 invokes the program for identifying time information from the voice information stored in the memory 105.
  • the program may also perform the following operations:
  • the date field in the intermediate data is assigned the same value as the current time.
  • the apparatus for identifying time information from the voice information and the terminal in the present embodiment described in FIG. 1 put the time features that may appear in the text information of the voice information into the configuration file of the corresponding category, and configure the configuration files of each category with Each field in the intermediate data corresponds one-to-one.
  • the matching time feature is searched from the configuration file, and the time character corresponding to the matching time feature is obtained, and the corresponding time character is assigned to the intermediate data.
  • the corresponding field in the field and generate the exact time data according to the content of the field, so that all the key time features to be matched are stored in the respective configuration files, and the identification program only needs to extract the time feature corresponding from the configuration file.
  • the time character, and the exact time data obtained through the intermediate file conversion makes the program code more readable, and effectively avoids the length of the program code.
  • the time feature needs to be modified or added, only the configuration file needs to be modified, and no modification is needed.
  • Re-encoding the program code to simplify the process of identifying the program, for the program code Maintenance is provided for convenience.
  • FIG. 2 is a flowchart of a first embodiment of a method for identifying time information from voice information according to the present application.
  • the method for identifying time information from voice information mentioned in this embodiment includes the following steps:
  • Step S10 collecting voice information, and identifying voice information to obtain text information corresponding to the voice information
  • the terminal carrier in this embodiment may be a mobile phone, a pad, a computer connected to a microphone, etc.
  • the terminal collects the voice information of the user through a voice collection device such as a microphone or a microphone.
  • the time information may be information for describing time, which is encoded in the collected voice information and encoded in a voice format.
  • the voice information is obtained as an operable control command, and the terminal performs a corresponding operation according to the control instruction.
  • the terminal system includes a common speech recognition program, and the speech information is obtained by the speech recognition program, and then the content of the speech information is converted into a standard format by the manner of the embodiment, so as to generate an operable Control instructions.
  • the content included in the voice information is not limited.
  • the time information in the voice information is taken as an example.
  • the identification and conversion of other content may also adopt the solution in this embodiment.
  • Step S20 identifying at least one time feature that matches the text information, and searching for at least one configuration file according to the matched time feature to obtain a corresponding time character;
  • the time feature may be information for describing time encoded in a text format, which is included in the text information.
  • This embodiment is further provided with a plurality of configuration files, each configuration file corresponding to one category.
  • the configuration file of the week includes the time characteristics of the day of the week, the week, and the day of the week
  • the configuration file of the date includes the number of days and days.
  • the month's profile includes a few months of time characteristics
  • the hour's profile includes the time characteristics of the time to wait.
  • There are various ways to identify the time feature in the text information In this embodiment, the text information content of the voice information is matched with the time feature in each configuration file to obtain a matching time feature.
  • the text information of the voice information is "Remind me to a meeting at 2:30 pm on May 8”
  • the time feature "May” is matched in the month of the configuration file month
  • the time feature is matched in the date configuration file day” 8th
  • the morning/afternoon profile halftime matches the time feature "afternoon”
  • the hour profile in the hour matches the time feature "2 points”
  • the minute profile min matches the time feature "Half”.
  • time features can also be divided into absolute time features and relative time features. For example, 10:30, 10, January 1, the first five days of the first month, Friday, Father's Day, National Day, New Year's Eve, etc., will not change due to the current time, so they are all absolute time characteristics. After 30 minutes, next February 14th, next month 3rd, next Wednesday, and the next day, you need to use the current time as a reference point. If there is a change in the current time, the specific time corresponding to the time feature will also change. Relative time characteristics.
  • a plurality of time features of the classification are stored in the configuration file, each time feature corresponding to a time character, as shown in FIG.
  • “May” corresponds to the time character 5
  • “8th” corresponds to the time character 8.
  • the corresponding time characters are also the same, for example, in Figure 10, Saturday, Sunday, Sunday, Sunday, Sunday, Sunday, Sunday, Sunday, Sunday, Sunday, Sunday, these six times.
  • the meanings of the features are the same, and the corresponding time characters are 7.
  • the time features "point half" and "point 30" in the minute profile min have the same meaning, and the corresponding time symbols are all 30.
  • the time character may be information for describing time, which is encoded in a text format and included in the configuration file.
  • the time feature and the time symbol may be information that are both encoded in a text format. As can be seen from the above, the difference between the two may be that the two describe the time differently.
  • Step S30 assigning a corresponding time character to a field corresponding to the configuration file in which the matching time feature is located in the intermediate data
  • the embodiment further has intermediate data, and the intermediate data includes a plurality of fields, each of which corresponds to a configuration file.
  • the terminal assigns the corresponding time symbol of the time feature to the field corresponding to the configuration file in which the matching time feature is located.
  • the configuration file in which the time feature "May” is located is the month configuration file month
  • the corresponding field in the intermediate data is the month field month
  • the variable in the month field is m.
  • the unit is "month”
  • step S40 time data corresponding to the time information in the voice information is obtained according to the content of each field in the intermediate data.
  • the terminal obtains the exact time data according to the assigned fields in the intermediate data.
  • the time data obtained includes absolute time and relative time.
  • the absolute time includes several minutes, several months, etc. It can also include the date of the lunar calendar and holidays, such as the fifth day of May, the Mid-Autumn Festival, the Father's Day, etc.; the relative date is relative to An offset of the current time, for example, today is March 1, when the content obtained according to the assigned field in the intermediate data is five days later, then the current date is increased by 5 days, and the exact time obtained after the conversion is March. 6th.
  • the time features that may appear in the text information of the voice information are placed in the configuration file of the corresponding category, and the configuration files of each category are corresponding to the fields in the intermediate data, and when the time information in the voice information is recognized.
  • Finding the matching time feature from the configuration file, and obtaining the time character corresponding to the matching time feature assigning the corresponding time character to the corresponding field in the intermediate data, and generating exact time data according to the content of the field, such that All the key time features to be matched are stored in their respective configuration files.
  • the identification program only needs to extract the time character corresponding to the time feature from the configuration file, and obtain the exact time data through the intermediate file conversion, so that the program code can be It is more readable and avoids the length of the program code.
  • you need to modify or add time features you only need to modify the configuration file, without modifying and re-encoding the program code, simplifying the modification process of the recognition program and providing maintenance for the program code. Convenient.
  • FIG. 3 is a flowchart of a second embodiment of a method for identifying time information from voice information according to the present application.
  • the modification of the configuration file in the method for identifying time information from the voice information mentioned in this embodiment adopts the following steps:
  • Step S51 receiving a new time feature entered, and determining a category corresponding to the new time feature
  • Step S52 adding a new time feature to the configuration file of the corresponding category of the new time feature
  • Step S53 it is determined whether there is a time feature with the same meaning of the new time feature in the configuration file of the corresponding category; if yes, step S54 is performed; if not, step S55 is performed;
  • Step S54 setting a time symbol corresponding to the new time feature according to the time character of the same time feature
  • step S55 a new time character is set for the new time feature.
  • a new time feature when modifying or adding a new key feature of speech recognition, only a new time feature needs to be added in the configuration file of the corresponding category, and a corresponding time symbol is set for the new time feature. If a time feature having the same meaning as the new time feature is included in the configuration file, the time character corresponding to the new time feature is set according to the time character of the same time feature; otherwise, the new time character is set to the new time The corresponding time character of the feature.
  • the time characteristic data needs to be modified or added, only the configuration file needs to be modified, and the program code is not modified and re-encoded, which simplifies the modification process of the identification program, and provides convenience for the maintenance of the program code.
  • FIG. 4 is a flowchart of a third embodiment of a method for identifying time information from voice information according to the present application.
  • This embodiment includes the steps of the embodiment shown in FIG. 2, and step S20 further includes:
  • Step S21 searching for a time feature table in each configuration file to find at least one time feature that matches the text information
  • Step S22 Find a time symbol corresponding to the matched time feature in the time-staff table in the configuration file in which the matched time feature is located.
  • the configuration file in this embodiment includes two tables, as shown in FIG. 11, one is a time feature table, and the other is a time character table.
  • the time feature of the corresponding category of the configuration file is stored in the time feature.
  • a plurality of time symbols are preset in the time list, and each time feature in the time feature table is respectively mapped with a time character in the time list, and the number of time features may be less than or equal to the time character. The number, multiple time features may correspond to a time character.
  • the time feature is added, the added time feature is placed in the time feature table, and the original time feature with the same meaning as the newly added time feature is found in the time feature table.
  • the configuration file of the embodiment adopts the time feature table and the time symbol table respectively, the time feature and the time character are respectively stored, so that the matching search is more convenient.
  • the increase of the time feature and the time character are respectively performed in the respective tables, which is advantageous for respectively extending the time feature table and the time symbol table.
  • the embodiment shown in FIG. 10 needs to repeat the same time character multiple times in the configuration file, and in the embodiment shown in FIG. 11, it is not necessary to be in the configuration file.
  • FIG. 5 is a method for identifying time information from voice information according to the present application.
  • Step S61 when the AM/PM field is not assigned in the intermediate data that has been assigned, and the date field is not assigned or is the same as the current time, the current time is obtained;
  • Step S62 it is determined whether the time symbol of the time point field in the intermediate data has been greater than the current time point; if yes, step S63 is performed; if not, step S64 is performed;
  • Step S63 setting the AM/PM field of the intermediate data to be consistent with the time period of the current time
  • step S64 the AM/PM field of the intermediate data is set to a period different from the period of the current time. For example, if the current time period is AM, the intermediate data is set to the afternoon field.
  • the size of the matching time feature of the current time and the text information of the voice information can be compared to determine whether the time period in the voice information is consistent with the current time period. For example, the current time is 9:00 AM.
  • the matching time feature is “10 points”, which is greater than the current time, and the “10 points” in the default voice information is At 10 o'clock in the morning, the alarm clock is set to 10 am; if the text information of the voice message is "8 points remind me to meet”, the matching time feature "8 points" is less than the current time, then the "8 points" in the default voice information For 8 pm, the reminder alarm is set to 8 pm. In this way, the time conversion error caused by misunderstanding is avoided, which is beneficial to improve the accuracy of time recognition in the voice information.
  • FIG. 6 is a method for identifying time information from voice information according to the present application.
  • a flow chart of a fifth embodiment This embodiment includes the steps of the embodiment shown in FIG. 2, and before step S40, the method further includes:
  • Step S71 when the week field has been assigned in the intermediate data that has been assigned, and the current week/next week field is not assigned, the current time is obtained;
  • Step S72 it is determined whether the time of the week field in the intermediate data has been greater than the number of days in the current time; if yes, step S73 is performed; if not, step S74 is performed;
  • Step S74 setting the current week/next week field of the intermediate data to this week;
  • step S73 the current week/next week field of the intermediate data is set to the next week.
  • This embodiment corresponds to incomplete information, such as "When me reminded me on Wednesday” refers to this Wednesday or next Wednesday? At this time, priority needs to be recognized as the future time.
  • the size of the matching time feature of the text information of the current time and the voice information can be compared to determine whether the number of weeks in the voice information is consistent with the number of weeks in the current time. For example, the current day's number of weeks is Monday.
  • the matching time feature is “Wednesday”, which is greater than the current time, and the “Wednesday” in the default voice message is this week.
  • the reminder alarm is set to be Wednesday; if the current time of the week is Friday, the matching time feature "Wednesday" is less than the current time of the week, the "Wednesday” in the default voice message is Wednesday of the next week. Remind the alarm to be set to next Wednesday. In this way, the time conversion error caused by misunderstanding is avoided, which is beneficial to improve the accuracy of time recognition in the voice information.
  • FIG. 7 is a flowchart of a sixth embodiment of a method for identifying time information from voice information according to the present application.
  • This embodiment includes the steps of the embodiment shown in FIG. 2, and before step S40, the method further includes:
  • Step S81 when the matched time feature includes tomorrow, the current time is acquired;
  • step S82 when the current time is in the range from 0:00 am to the morning x point, where x is the preset time point, the date field in the intermediate data is assigned the same value as the current time.
  • the current time is 1 am on the 7th
  • the voice message “Make me a meeting at 10 o'clock tomorrow morning” refers to 10 am on the 7th or 10 am on the 8th.
  • the user can set a time range in advance, and the default time range is the morning time period, and the default time is the previous day in the early morning time period. For example, if the preset time point is 5 o'clock, the preset morning time period is from 0:00 am to 5:00 am, and the current time is 1 am (1 am, 1 am).
  • the voice information is " To remind me of the meeting at 10 o'clock tomorrow morning, it means 10 am on the 7th, reminding the alarm to be set to 10 am on the 7th. In this way, the time conversion error caused by misunderstanding is avoided, which is beneficial to improve the accuracy of time recognition in the voice information.
  • FIG. 8 is a schematic block diagram of a first embodiment of an apparatus for identifying time information from voice information according to the present application.
  • the device for identifying time information from the voice information mentioned in this embodiment includes:
  • the voice collection module 110 is configured to collect voice information, and identify voice information to obtain text information corresponding to the voice information.
  • the configuration query module 120 is configured to identify at least one time feature that matches the text information, and find at least one configuration file according to the matched time feature to obtain a corresponding time character;
  • the intermediate assignment module 130 is configured to assign a corresponding time character to a field in the intermediate data corresponding to the configuration file where the matched time feature is located;
  • the time conversion module 140 is configured to obtain time data corresponding to time information in the voice information according to content of each field in the intermediate data.
  • the terminal carrier in this embodiment may be a mobile phone, a pad, a computer connected to a microphone, etc.
  • the terminal collects the voice information of the user through a voice collection device such as a microphone or a microphone.
  • a voice collection device such as a microphone or a microphone.
  • the terminal system includes a common speech recognition program, and the speech information is obtained by the speech recognition program, and then the content of the speech information is converted into a standard format by the manner of the embodiment, so as to generate an operable Control instructions.
  • the content included in the voice information is not limited.
  • the time information in the voice information is taken as an example.
  • the identification and conversion of other content may also adopt the solution in this embodiment.
  • This embodiment is further provided with a plurality of configuration files, each configuration file corresponding to one category.
  • the configuration file of the week includes the time characteristics of the day of the week, the week, and the day of the week
  • the configuration file of the date includes the number of days and days.
  • the month's profile includes a few months of time characteristics
  • the hour's profile includes the time characteristics of the time to wait.
  • the text information content of the voice information is matched with the time feature in each configuration file to obtain a matching time feature.
  • the text information of the voice information is "Remind me to a meeting at 2:30 pm on May 8”
  • the time feature "May” is matched in the month of the configuration file month
  • the time feature is matched in the date configuration file day”
  • the AM/PM profile halftime matches the time feature "PM”
  • the hour profile hour matches the time feature "2 points”
  • the minute profile min matches the time feature "Point Half”.
  • time features can also be divided into absolute time features and relative time features. For example, 10:30, 10, January 1, the first five days of the first month, Friday, Father's Day, National Day, New Year's Eve, etc., will not change due to the current time, so they are all absolute time characteristics. After 30 minutes, next February 14th, next month 3rd, next Wednesday, and the next day, you need to use the current time as a reference point. If there is a change in the current time, the specific time corresponding to the time feature will also change. Relative time characteristics.
  • a plurality of time features of the classification are stored in the configuration file, each time feature corresponding to a time character, as shown in FIG.
  • “May” corresponds to the time character 5
  • in the configuration file of the date "8th” corresponds to the time character 8.
  • the corresponding time symbols are also the same, for example, in Figure 10, on the seventh day, Sunday, Sunday, Sunday, Sunday, Sunday, Sunday, Sunday, Sunday, Sunday, Sunday, Sunday, Sunday, these six time features have the same meaning, corresponding The time character is 7.
  • the time features "point half" and "point 30" in the minute profile min have the same meaning, and the corresponding time symbols are all 30.
  • the embodiment further has intermediate data, and the intermediate data includes a plurality of fields, each of which corresponds to a configuration file.
  • the terminal assigns the corresponding time symbol of the time feature to the field corresponding to the configuration file in which the matching time feature is located.
  • the configuration file in which the time feature "May” is located is the month configuration file month
  • the corresponding field in the intermediate data is the month field month
  • the variable in the month field is m
  • the unit is "month”
  • the time feature is "5".
  • the terminal obtains the exact time data according to the assigned fields in the intermediate data.
  • the time data obtained includes absolute time and relative time.
  • the absolute time includes several minutes, several months, etc. It can also include the date of the lunar calendar and holidays, such as the fifth day of May, the Mid-Autumn Festival, the Father's Day, etc.; the relative date is relative to An offset of the current time, for example, today is March 1, when the content obtained according to the assigned field in the intermediate data is five days later, then the current date is increased by 5 days, and the exact time obtained after the conversion is March. 6th.
  • the time features that may appear in the text information of the voice information are placed in the configuration file of the corresponding category, and the configuration files of each category are corresponding to the fields in the intermediate data, and when the time information in the voice information is recognized.
  • Finding the matching time feature from the configuration file, and obtaining the time character corresponding to the matching time feature assigning the corresponding time character to the corresponding field in the intermediate data, and generating exact time data according to the content of the field, such that All the key time features to be matched are stored in their respective configuration files.
  • the recognition program only needs to extract the time character corresponding to the time feature from the configuration file and convert it through the intermediate file. Obtain the exact time data, make the program code more readable, and effectively avoid the length of the program code.
  • you need to modify or add the time feature data you only need to modify the configuration file, without modifying and re-encoding the program code, simplifying Identifying the modification process of the program provides convenience for the maintenance of the program code.
  • FIG. 9 is a schematic block diagram of a second embodiment of an apparatus for identifying time information from voice information according to the present application.
  • the embodiment includes the module of the embodiment shown in FIG. 8, and further includes a configuration modification module 150, configured to:
  • the time character corresponding to the new time feature is set according to the time character of the same time feature
  • a new time feature when modifying or adding a new key feature of speech recognition, only a new time feature needs to be added in the configuration file of the corresponding category, and a corresponding time symbol is set for the new time feature. If a time feature having the same meaning as the new time feature is included in the configuration file, the time character corresponding to the new time feature is set according to the time character of the same time feature; otherwise, the new time character is set to the new time The corresponding time character of the feature.
  • the time characteristic data needs to be modified or added, only the configuration file needs to be modified, and the program code is not modified and re-encoded, which simplifies the modification process of the identification program, and provides convenience for the maintenance of the program code.
  • configuration query module 120 is further configured to:
  • the time character corresponding to the matched time feature is found in the time-staff table in the configuration file in which the matching time feature is located.
  • the configuration file in this embodiment includes two tables, as shown in FIG. 11, one is a time feature table, and the other is a time character table.
  • the time feature of the corresponding category of the configuration file is stored in the time feature.
  • a plurality of time symbols are preset in the time list, and each time feature in the time feature table is respectively mapped with a time character in the time list, and the number of time features may be less than or equal to the time character. The number, multiple time features may correspond to a time character.
  • the time feature is added, the added time feature is placed in the time feature table, and the original time feature with the same meaning as the newly added time feature is found in the time feature table.
  • the configuration file of the embodiment adopts the time feature table and the time symbol table respectively, the time feature and the time character are respectively stored, so that the matching search is more convenient.
  • the increase of the time feature and the time character are respectively performed in the respective tables, which is advantageous for respectively extending the time feature table and the time symbol table.
  • the embodiment shown in FIG. 10 needs to repeat the same time character multiple times in the configuration file, and in the embodiment shown in FIG. 11, it is not necessary to be in the configuration file.
  • intermediate assignment module 130 is further configured to:
  • the current time is obtained when the AM/PM field is not assigned in the intermediate data that has been assigned, and the date field is not assigned or the same date as the current time;
  • the AM/PM field of the intermediate data is set to coincide with the current time period
  • the AM/PM field of the intermediate data is set to be opposite to the current time period.
  • the size of the matching time feature of the current time and the text information of the voice information can be compared to determine whether the time period in the voice information is consistent with the current time period. For example, the current time is 9:00 AM.
  • the matching time feature is “10 points”, which is greater than the current time, and the “10 points” in the default voice information is At 10 o'clock in the morning, the alarm clock is set to 10 am; if the text information of the voice message is "8 points remind me to meet”, the matching time feature "8 points" is less than the current time, then the "8 points" in the default voice information For 8 pm, the reminder alarm is set to 8 pm. In this way, the time conversion error caused by misunderstanding is avoided, which is beneficial to improve the accuracy of time recognition in the voice information.
  • intermediate assignment module 130 is further configured to:
  • This embodiment corresponds to incomplete information, such as "When me reminded me on Wednesday” refers to this Wednesday or next Wednesday? At this time, priority needs to be recognized as the future time.
  • the size of the matching time feature of the text information of the current time and the voice information can be compared to determine whether the number of weeks in the voice information is consistent with the number of weeks in the current time. For example, the current day's number of weeks is Monday.
  • the matching time feature is “Wednesday”, which is greater than the current time, and the “Wednesday” in the default voice message is this week.
  • the reminder alarm is set to be Wednesday; if the current time of the week is Friday, the matching time feature "Wednesday" is less than the current time of the week, the "Wednesday” in the default voice message is Wednesday of the next week. Remind the alarm to be set to next Wednesday. In this way, the time conversion error caused by misunderstanding is avoided, which is beneficial to improve the accuracy of time recognition in the voice information.
  • intermediate assignment module 130 is further configured to:
  • the date field in the intermediate data is assigned the same value as the current time.
  • the current time is 1 am on the 7th
  • the voice message “Make me a meeting at 10 o'clock tomorrow morning” refers to 10 am on the 7th or 10 am on the 8th.
  • the user can set a time range in advance, and the default time range is the morning time period, and the default time is the previous day in the early morning time period. For example, if the preset time point is 5 o'clock, the preset morning time period is from 0:00 am to 5:00 am, and the current time is 1 am (1 am, 1 am).
  • the voice information is " To remind me of the meeting at 10 o'clock tomorrow morning, it means 10 am on the 7th, reminding the alarm to be set to 10 am on the 7th. In this way, time conversion errors caused by misunderstandings are avoided. It is beneficial to improve the accuracy of time recognition in voice information.
  • the foregoing embodiment method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be through hardware, but in many cases, the former is better.
  • Implementation Based on such understanding, the technical solution of the present application, which is essential or contributes to the prior art, may be embodied in the form of a software product stored in a storage medium (such as ROM/RAM, disk,
  • the optical disc includes a number of instructions for causing a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) to perform the methods described in various embodiments of the present application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Multimedia (AREA)
  • Human Computer Interaction (AREA)
  • Acoustics & Sound (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • General Health & Medical Sciences (AREA)
  • Telephonic Communication Services (AREA)
  • Telephone Function (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种从语音信息中识别时间信息的方法和装置，该方法包括：采集语音信息，识别包含时间信息的语音信息，获得文本信息(S10)；识别与文本信息匹配的至少一个时间特征，并根据匹配的时间特征查找至少一个配置文件以获得对应的时间符(S20)；将时间符赋值给中间数据中与时间特征所在的配置文件对应的字段(S30)；根据中间数据中各字段的内容，获得时间数据(S40)。该方法将时间特征存放在配置文件中，使程序代码可读性更强，避免了代码冗长，在需要修改或添加时间特征数据时，只需要修改配置文件，无须改动和重新编码程序代码，简化了识别程序的修改过程，为程序代码的维护提供了方便。

Description

从语音信息中识别时间信息的方法和装置
本申请要求于2015年7月31日提交中国专利局、申请号为201510465853.7、发明名称为“从自然语音信息中识别时间信息的方法和装置”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及到语音识别技术领域，特别涉及到从语音信息中识别时间信息的方法和装置。
背景技术
随着语音识别技术的发展，将语音识别结合到终端应用程序中，在终端中通过语音交互实现搜索查询或操作功能，例如查询天气、航班动态、搜索周边美食、拨打电话、设置提醒、调整闹钟、播放音乐等。在用户与终端进行语音交互时，语音指令中经常会携带时间信息，例如5月5日早上10点提醒开会、查询下周一深圳到北京的航班、设置明天早上8点的闹钟等。
发明内容
本申请实施例提供一种从语音信息中识别时间信息的方法和装置，在修改或添加语音指令的识别关键词时，无须改动或重新编译程序代码，可操作性强。
本申请实施例提出一种从语音信息中识别时间信息的方法，包括步骤：
采集语音信息，并识别包含时间信息的语音信息获得对应于所述语 音信息的文本信息；
识别与所述文本信息匹配的至少一个时间特征，并根据所述匹配的时间特征查找至少一个配置文件以获得对应的时间符；
将所述对应的时间符赋值给中间数据中与所述匹配的时间特征所在的配置文件对应的字段；所述中间数据包括多个字段，每个字段对应于一个所述配置文件；
根据所述中间数据中各字段的内容，获得与所述语音信息中的所述时间信息对应的时间数据。
本申请实施例还提出一种从语音信息中识别时间信息的装置，包括：
语音采集模块，用于采集语音信息，并识别包含时间信息的语音信息获得对应于所述语音信息的文本信息；
配置查询模块，用于识别与所述文本信息匹配的至少一个时间特征，并根据所述匹配的时间特征查找至少一个配置文件以获得对应的时间符；
中间赋值模块，用于将所述对应的时间符赋值给中间数据中与所述匹配的时间特征所在的配置文件对应的字段；所述中间数据包括多个字段，每个字段对应于一个所述配置文件；
时间转换模块，用于根据所述中间数据中各字段的内容，获得与所述语音信息中的所述时间信息对应的时间数据。
本申请实施例将语音信息中可能出现的时间特征放入对应类别的配置文件中，将各类别的配置文件与中间数据中的各字段一一对应，在识别语音信息中的时间信息时，从配置文件中查找匹配的时间特征，并 获得匹配时间特征对应的时间符，将该对应的时间符赋值给中间数据中的对应字段，并根据该字段的内容生成确切的时间数据，如此一来，所有待匹配的关键时间特征都存放在各自的配置文件中，识别程序只需要从配置文件中提取时间特征对应的时间符，并通过中间文件转换获得确切的时间数据，使程序代码的可读性更强，且有效避免了程序代码冗长，在需要修改或添加时间特征时，只需要修改配置文件，无须改动和重新编码程序代码，简化了识别程序的修改过程，为程序代码的维护提供了方便。
附图说明
图1为本申请实施例中从语音信息中识别时间信息的装置所在终端的总线图；
图2为本申请从语音信息中识别时间信息的方法的第一实施例的流程图；
图3为本申请从语音信息中识别时间信息的方法的第二实施例的流程图；
图4为本申请从语音信息中识别时间信息的方法的第三实施例的流程图；
图5为本申请从语音信息中识别时间信息的方法的第四实施例的流程图；
图6为本申请从语音信息中识别时间信息的方法的第五实施例的流程图；
图7为本申请从语音信息中识别时间信息的方法的第六实施例的流程图；
图8为本申请从语音信息中识别时间信息的装置的第一实施例的模 块示意图；
图9为本申请从语音信息中识别时间信息的装置的第二实施例的模块示意图；
图10为本申请中配置文件的第一实施例的示意图；
图11为本申请中配置文件的第二实施例的示意图。
本申请目的的实现、功能特点及优点将结合实施例，参照附图做进一步说明。
具体实施方式
应当理解，此处所描述的具体实施例仅仅用以解释本申请，并不用于限定本申请。
在本申请一实施例中，采用硬编码方式对于语音指令中的时间信息进行识别。具体是，在程序代码中编码各种关键词，在识别到语音指令中的关键词后，将其转换为对应的时间参数，生成对应的时间。但是，这种硬编码方式需要将所有的时间关键词都添加到代码中。
参照图1，图1为本申请实施例中从语音信息中识别时间信息的装置所在终端的总线图。该终端可以是手机、pad、电脑等，可以包括：至少一个处理器101，例如CPU，至少一个网络接口104，用户接口103，存储器105，至少一个通信总线102。其中，通信总线102用于实现这些组件之间的连接通信。其中，用户接口103可以包括显示屏(Display)、键盘(Keyboard)，还可以包括标准的有线接口、无线接口。网络接口104可以包括标准的有线接口、无线接口(如WIFI接口)。存储器105可以是高速RAM存储器，也可以是非不稳定的存储器(non-volatile memory)，例如至少一个磁盘存储器。存储器105还可以是至少一个位于远离前述处理器101的存储装置。作为一种计算机存储介质的存储器 105中可以包括操作系统、网络通信模块、用户接口模块以及从语音信息中识别时间信息程序。在本发明实施例中，语音信息可以为自然语音信息，例如从人的表达的语言中识别出的语音信息。
在图1所示的从语音信息中识别时间信息的装置所在终端中，网络接口104主要用于连接服务器或其他终端，与服务器或其他终端进行数据通信；而用户接口103主要用于接收用户指令，并与用户进行交互；而处理器101可以用于调用存储器105中存储的从语音信息中识别时间信息程序，并执行以下操作：
采用用户接口103采集语音信息，并识别包含时间信息的语音信息获得对应于语音信息的文本信息；
识别与文本信息匹配的至少一个时间特征，并根据匹配的时间特征查找至少一个配置文件以获得对应的时间符；
将对应的时间符赋值给中间数据中与匹配的时间特征所在的配置文件对应的字段；中间数据包括多个字段，每个字段对应于一个配置文件；
根据中间数据中各字段的内容，获得与语音信息中的时间信息对应的时间数据。
在本发明实施例中，中间数据中各字段可以为：与上述各配置文件对应的字段。
在一个实施例中，处理器101调用存储器105中存储的从语音信息中识别时间信息程序还可以执行以下操作：
采用用户接口103接收录入的新的时间特征，确定新的时间特征对应的类别；
将新的时间特征添加到新的时间特征对应类别的配置文件中；
判断在对应类别的配置文件中是否存在与新的时间特征含义相同 的时间特征；
如果是，则根据含义相同的时间特征的时间符设置新的时间特征对应的时间符；
如果否，则对所述新的时间特征设置新的时间符。
在一个实施例中，处理器101调用存储器105中存储的从语音信息中识别时间信息程序还可以执行以下操作：
在各个配置文件中的时间特征表中查找与文本信息匹配的至少一个时间特征；
在匹配的时间特征所在的配置文件中的时间符表中查找与匹配的时间特征对应的时间符。
在一个实施例中，处理器101调用存储器105中存储的从语音信息中识别时间信息程序还可以执行以下操作：
当已完成赋值的中间数据中上午/下午字段未赋值，且日期字段未赋值或与当前时间的日期相同时，获取当前时间；
判断已完成赋值的中间数据中时间点字段的时间符是否大于当前时间的时间点；
如果是，则将中间数据的上午/下午字段设置为与当前时间的时段一致；
如果否，则将中间数据的上午/下午字段设置为与当前时间的时段相反。
在一个实施例中，处理器101调用存储器105中存储的从语音信息中识别时间信息程序还可以执行以下操作：
当已完成赋值的中间数据中星期字段已赋值，且本周/下周字段未赋值时，获取当前时间；
判断已完成赋值的中间数据中星期字段的时间符是否大于当前时 间的星期数；
如果是，则将中间数据的本周/下周字段设置为本周；
如果否，则将中间数据的本周/下周字段设置为下周。
在一个实施例中，处理器101调用存储器105中存储的从语音信息中识别时间信息程序还可以执行以下操作：
当匹配的时间特征中包括明天时，获取当前时间；
在当前时间在上午0点至上午x点范围内时，其中x为预设时间点，则将中间数据中日期字段赋值为与当前时间的日期相同。
本实施例图1所描述的从语音信息中识别时间信息的装置及其所在终端，将语音信息的文本信息中可能出现的时间特征放入对应类别的配置文件中，将各类别的配置文件与中间数据中的各字段一一对应，在识别语音信息中的时间信息时，从配置文件中查找匹配的时间特征，并获得匹配时间特征对应的时间符，将该对应的时间符赋值给中间数据中的对应字段，并根据该字段的内容生成确切的时间数据，如此一来，所有待匹配的关键时间特征都存放在各自的配置文件中，识别程序只需要从配置文件中提取时间特征对应的时间符，并通过中间文件转换获得确切的时间数据，使程序代码的可读性更强，且有效避免了程序代码冗长，在需要修改或添加时间特征时，只需要修改配置文件，无须改动和重新编码程序代码，简化了识别程序的修改过程，为程序代码的维护提供了方便。
如图2所示，图2为本申请从语音信息中识别时间信息的方法的第一实施例的流程图。本实施例提到的从语音信息中识别时间信息的方法，包括步骤：
步骤S10，采集语音信息，并识别语音信息获得对应于所述语音信息的文本信息；
本实施例中的终端载体可以是手机、pad、连接有麦克风的电脑等，当用户启动终端的语音采集功能时，终端通过话筒、麦克风等语音采集设备采集用户的语音信息。在本发明一实施例中，时间信息可以为采集到的语音信息中包含的，以语音格式进行编码的用于描述时间的信息。本实施例不止是对语音信息的内容进行识别，还要将语音信息获得可操作的控制指令，由终端根据该控制指令执行相应操作。因此，终端系统中包括有普通的语音识别程序，通过语音识别程序将语音信息获得对应的文本信息，然后再由本实施例的方式来将语音信息的内容转换成标准的格式，以便于生成可操作的控制指令。该语音信息包括的内容并不限定，本实施例以识别语音信息中的时间信息为例，其他内容的识别与转换也可采用本实施例的方案。
步骤S20，识别与文本信息匹配的至少一个时间特征，并根据匹配的时间特征查找至少一个配置文件以获得对应的时间符；
在本发明一实施例中，时间特征可以是包含在文本信息中的，以文本格式进行编码的用于描述时间的信息。本实施例还设有多个配置文件，每个配置文件对应一个分类，例如，星期的配置文件包括星期几、礼拜几、星期几的时间特征，日期的配置文件包括几号、几日的时间特征，月份的配置文件包括几月的时间特征，小时的配置文件包括几点钟的时间特征等待。在文本信息中识别时间特征的方式有多种，本实施例采用的是将语音信息的文本信息内容与各个配置文件中的时间特征进行匹配，获得匹配的时间特征。例如，语音信息的文本信息为“5月8日下午2点半提醒我开会”，则在月份的配置文件month中匹配到时间特征“5月”，日期的配置文件day中匹配到时间特征“8日”，上午/下午的配置文件halfday中匹配到时间特征“下午”，小时的配置文件hour中匹配到时间特征“2点”，分钟配置文件min中匹配到时间特征 “点半”。
此外，时间特征还可以分为绝对时间特征和相对时间特征。例如，十点半、十号、一月一号、正月初五、星期五、父亲节、国庆节、除夕等，不会因当前时间不同而产生变化，因此都是绝对时间特征。30分钟后、明年2月14号、下个月3号、下周三、后天等，需要以当前时间作为参照点，当前时间若有变化，则该时间特征对应的具体时间也会变化，因此是相对时间特征。
在配置文件中存储有该分类的多个时间特征，每一个时间特征对应于一个时间符，如图10所示。例如月份的配置文件中“5月”对应时间符5；日期的配置文件中“8日”对应时间符8。对于含义相同的多个时间特征，对应的时间符也相同，例如图10中，周七、周日、周天、星期七、星期日、星期天、礼拜七、礼拜日、礼拜天，这六个时间特征的含义相同，对应时间符都为7。前述实施例中，分钟配置文件min中的时间特征“点半”、“点30”的含义相同，对应时间符都为30。在本发明一实施例中，时间符可以是包含在配置文件中的，以文本格式进行编码的用于描述时间的信息。时间特征和时间符可以是均采用文本格式进行编码的信息，如上可见，两者的区别可在于两者描述时间的方式不同。
步骤S30，将对应的时间符赋值给中间数据中与匹配的时间特征所在的配置文件对应的字段；
本实施例还设有中间数据，中间数据中包括有多个字段，每一个字段对应于一个配置文件。终端在查找到文本信息匹配的时间特征后，将该时间特征的对应时间符赋值给匹配时间特征所在配置文件对应的字段中。例如，时间特征“5月”所在的配置文件为月份配置文件month，在中间数据中对应的字段为月份字段month，该月份字段中的变量为m， 单位为“月”，将时间特征“5月”对应的时间符5赋值给月份字段的变量，即令m＝5，赋值后该字段为“5月”。
步骤S40，根据中间数据中各字段的内容，获得与语音信息中的时间信息对应的时间数据。
将本次语音信息的文本信息中所有匹配时间特征对应的时间符都赋值给中间数据的对应字段后，终端根据中间数据中已赋值的字段，获得确切的时间数据。获得的时间数据包括绝对时间和相对时间，绝对时间包括几点几分、几月几号等，还可包括农历日期和节假日，例如五月初五、中秋节、父亲节等；相对日期是相对于当前时间的一个偏移量，例如今天是3月1日，根据中间数据中已赋值的字段获得的内容为五天后，则在当前日期往后增加5天，转换后获得的确切时间为3月6日。
本实施例将语音信息的文本信息中可能出现的时间特征放入对应类别的配置文件中，将各类别的配置文件与中间数据中的各字段一一对应，在识别语音信息中的时间信息时，从配置文件中查找匹配的时间特征，并获得匹配时间特征对应的时间符，将该对应的时间符赋值给中间数据中的对应字段，并根据该字段的内容生成确切的时间数据，如此一来，所有待匹配的关键时间特征都存放在各自的配置文件中，识别程序只需要从配置文件中提取时间特征对应的时间符，并通过中间文件转换获得确切的时间数据，使程序代码的可读性更强，且有效避免了程序代码冗长，在需要修改或添加时间特征时，只需要修改配置文件，无须改动和重新编码程序代码，简化了识别程序的修改过程，为程序代码的维护提供了方便。
如图3所示，图3为本申请从语音信息中识别时间信息的方法的第二实施例的流程图。本实施例提到的从语音信息中识别时间信息的方法中配置文件的修改采用以下步骤：
步骤S51，接收录入的新的时间特征，确定新的时间特征对应的类别；
步骤S52，将新的时间特征添加到新的时间特征对应类别的配置文件中；
步骤S53，判断在对应类别的配置文件中是否存在与新的时间特征含义相同的时间特征；如果是，则执行步骤S54；如果否，则执行步骤S55；
步骤S54，根据含义相同的时间特征的时间符设置新的时间特征对应的时间符；
步骤S55，对新的时间特征设置新的时间符。
本实施例在修改或增加新的语音识别的关键时间特征时，只需要在对应类别的配置文件中添加新的时间特征，并对新的时间特征设置对应时间符。如果在配置文件中包括有与新的时间特征含义相同的时间特征，则根据含义相同的时间特征的时间符设置新的时间特征对应的时间符；否则，将新的时间符设置为新的时间特征的对应时间符。本实施例在需要修改或添加时间特征数据时，只需要修改配置文件，无须改动和重新编码程序代码，简化了识别程序的修改过程，为程序代码的维护提供了方便。
如图4所示，图4为本申请从语音信息中识别时间信息的方法的第三实施例的流程图。本实施例包括图2所示实施例的步骤，步骤S20还包括：
步骤S21，在各个配置文件中的时间特征表中查找与文本信息匹配的至少一个时间特征；
步骤S22，在匹配的时间特征所在的配置文件中的时间符表中查找与匹配的时间特征对应的时间符。
本实施例中的配置文件包括两个表，如图11所示，一个是时间特征表，另一个是时间符表，在建立配置文件时，将该配置文件对应类别的时间特征存储在时间特征表中，而在时间符表中预先设置有多个时间符，将时间特征表中的各个时间特征分别与时间符表中的时间符建立映射关系，时间特征的数量可小于或等于时间符的数量，多个时间特征可对应于一个时间符。在新增时间特征时，将新增的时间特征放置于时间特征表中，并查找该时间特征表中是否有与新增的时间特征含义相同的原有时间特征。如果有，则在时间符表中查找与该含义相同的原有时间特征匹配的时间符，并在查找到的特征符与新增的时间特征之间建立新的映射关系。如果该时间特征表中没有与新增的时间特征含义相同的原有时间特征，则在时间符表中新建一新的时间符，并将新的时间符与新增的时间特征建立映射关系。在查找匹配的时间特征时，在各个配置文件的时间特征表中查找与文本信息匹配的时间特征，然后在该匹配的配置文件的时间符表中查找与匹配的时间特征对应的时间符即可。
由于本实施例的配置文件分别采用了时间特征表和时间符表，将时间特征和时间符分别存储，使匹配查找更加方便。同时，时间特征和时间符的增加分别在各自所在的表中进行，有利于分别扩展时间特征表和时间符表。此外，在多个时间特征对应于一个时间符时，图10所示实施例中需要将同一个时间符在配置文件中重复的出现多次，而图11所示实施例中，无须在配置文件中写入过多的重复数据，只需要在两个表之间建立映射关系即可，无论是时间特征还是时间符，在配置文件中都只出现一次，有利于减少配置文件中的数据量，避免配置文件占用过多存储空间。
如图5所示，图5为本申请从语音信息中识别时间信息的方法的第 四实施例的流程图。本实施例包括图2所示实施例的步骤，步骤S40之前还包括：
步骤S61，当已完成赋值的中间数据中上午/下午字段未赋值，且日期字段未赋值或与当前时间的日期相同时，获取当前时间；
步骤S62，判断已完成赋值的中间数据中时间点字段的时间符是否大于当前时间的时间点；如果是，则执行步骤S63；如果否，则执行步骤S64；
步骤S63，将中间数据的上午/下午字段设置为与当前时间的时段一致；
步骤S64，将中间数据的上午/下午字段设置为与当前时间的时段不同的时段。例如，当前时间所属的时段为上午，则将中间数据设置为下午字段。
本实施例在采集的语音信息中，常常会遇到一些不完整的信息，例如“10点提醒我开会”是指的上午十点还是下午十点？这时候则需要优先识别为未来的时间。此时可比较当前时间与语音信息的文本信息的匹配时间特征的大小来判断语音信息中的时段与当前时间的时段是否一致。例如，当前时间为上午9点，当语音信息的文本信息为“10点提醒我开会”时，匹配的时间特征为“10点”，大于当前时间，则默认语音信息中的“10点”为上午10点，提醒闹钟设置为上午10点；如果当语音信息的文本信息为“8点提醒我开会”，匹配时间特征“8点”，小于当前时间，则默认语音信息中的“8点”为下午8点，提醒闹钟设置为下午8点。如此一来，避免了因误解造成时间转换错误，有利于提高语音信息中时间识别的准确性。
如图6所示，图6为本申请从语音信息中识别时间信息的方法的第 五实施例的流程图。本实施例包括图2所示实施例的步骤，步骤S40之前还包括：
步骤S71，当已完成赋值的中间数据中星期字段已赋值，且本周/下周字段未赋值时，获取当前时间；
步骤S72，判断已完成赋值的中间数据中星期字段的时间符是否大于当前时间的星期数；如果是，则执行步骤S73；如果否，则执行步骤S74；
步骤S74，将中间数据的本周/下周字段设置为本周；
步骤S73，将中间数据的本周/下周字段设置为下周。
本实施例对应不完整的信息，例如“周三提醒我开会”是指的本周三还是下周三？这时候也需要优先识别为未来的时间。此时可比较当前时间与语音信息的文本信息的匹配时间特征的大小来判断语音信息中的星期数与当前时间的星期数是否一致。例如，当前时间的星期数为周一，当语音信息的文本信息为“周三提醒我开会”时，匹配的时间特征为“周三”，大于当前时间，则默认语音信息中的“周三”为本周的周三，提醒闹钟设置为本周三；如果当前时间的星期数为周五，匹配时间特征“周三”，小于当前时间的星期数，则默认语音信息中的“周三”为下周的周三，提醒闹钟设置为下周三。如此一来，避免了因误解造成时间转换错误，有利于提高语音信息中时间识别的准确性。
如图7所示，图7为本申请从语音信息中识别时间信息的方法的第六实施例的流程图。本实施例包括图2所示实施例的步骤，步骤S40之前还包括：
步骤S81，当匹配的时间特征中包括明天时，获取当前时间；
步骤S82，在当前时间在上午0点至上午x点范围内时，其中x为预设时间点，则将中间数据中日期字段赋值为与当前时间的日期相同。
本实施例在采集的语音信息中，还会遇到一种特殊的情况，人们往往在凌晨时间段认为是前一天，但实际上凌晨时间段已经是第二天。例如，当前时间是7号凌晨1点，语音信息“明天早上10点提醒我开会”是指的7号上午十点还是8号上午十点？这时候可预先由用户设置一个时间范围，默认该时间范围内为凌晨时间段，在该凌晨时间段内默认为是前一天。例如，预设时间点为5点，则预设的凌晨时间段为上午0点至上午5点，当前时间是7号上午1点(即凌晨1点)，在此范围内，则语音信息“明天早上10点提醒我开会”是指7号上午十点，提醒闹钟设置为7号上午十点。如此一来，避免了因误解造成时间转换错误，有利于提高语音信息中时间识别的准确性。
如图8所示，图8为本申请从语音信息中识别时间信息的装置的第一实施例的模块示意图。本实施例提到的从语音信息中识别时间信息的装置，包括：
语音采集模块110，用于采集语音信息，并识别语音信息获得对应于所述语音信息的文本信息；
配置查询模块120，用于识别与文本信息匹配的至少一个时间特征，并根据匹配的时间特征查找至少一个配置文件以获得对应的时间符；
中间赋值模块130，用于将对应的时间符赋值给中间数据中与匹配的时间特征所在的配置文件对应的字段；
时间转换模块140，用于根据中间数据中各字段的内容，获得与所述语音信息中的时间信息对应的时间数据。
本实施例中的终端载体可以是手机、pad、连接有麦克风的电脑等，当用户启动终端的语音采集功能时，终端通过话筒、麦克风等语音采集设备采集用户的语音信息。本实施例不止是对语音信息的内容进行识别，还要将语音信息获得可操作的控制指令，由终端根据该控制指令执 行相应操作。因此，终端系统中包括有普通的语音识别程序，通过语音识别程序将语音信息获得对应的文本信息，然后再由本实施例的方式来将语音信息的内容转换成标准的格式，以便于生成可操作的控制指令。该语音信息包括的内容并不限定，本实施例以识别语音信息中的时间信息为例，其他内容的识别与转换也可采用本实施例的方案。
本实施例还设有多个配置文件，每个配置文件对应一个分类，例如，星期的配置文件包括星期几、礼拜几、星期几的时间特征，日期的配置文件包括几号、几日的时间特征，月份的配置文件包括几月的时间特征，小时的配置文件包括几点钟的时间特征等待。在文本信息中识别时间特征的方式有多种，本实施例采用的是将语音信息的文本信息内容与各个配置文件中的时间特征进行匹配，获得匹配的时间特征。例如，语音信息的文本信息为“5月8日下午2点半提醒我开会”，则在月份的配置文件month中匹配到时间特征“5月”，日期的配置文件day中匹配到时间特征“8日”，上午/下午的配置文件halfday中匹配到时间特征“下午”，小时的配置文件hour中匹配到时间特征“2点”，分钟配置文件min中匹配到时间特征“点半”。
此外，时间特征还可以分为绝对时间特征和相对时间特征。例如，十点半、十号、一月一号、正月初五、星期五、父亲节、国庆节、除夕等，不会因当前时间不同而产生变化，因此都是绝对时间特征。30分钟后、明年2月14号、下个月3号、下周三、后天等，需要以当前时间作为参照点，当前时间若有变化，则该时间特征对应的具体时间也会变化，因此是相对时间特征。
在配置文件中存储有该分类的多个时间特征，每一个时间特征对应于一个时间符，如图10所示。例如月份的配置文件中“5月”对应时间符5；日期的配置文件中“8日”对应时间符8。对于含义相同的多个 时间特征，对应的时间符也相同，例如图10中，周七、周日、周天、星期七、星期日、星期天、礼拜七、礼拜日、礼拜天，这六个时间特征的含义相同，对应时间符都为7。前述实施例中，分钟配置文件min中的时间特征“点半”、“点30”的含义相同，对应时间符都为30。
本实施例还设有中间数据，中间数据中包括有多个字段，每一个字段对应于一个配置文件。终端在查找到文本信息匹配的时间特征后，将该时间特征的对应时间符赋值给匹配时间特征所在配置文件对应的字段中。例如，时间特征“5月”所在的配置文件为月份配置文件month，在中间数据中对应的字段为月份字段month，该月份字段中的变量为m，单位为“月”，将时间特征“5月”对应的时间符5赋值给月份字段的变量，即令m＝5，赋值后该字段为“5月”。
将本次语音信息的文本信息中所有匹配时间特征对应的时间符都赋值给中间数据的对应字段后，终端根据中间数据中已赋值的字段，获得确切的时间数据。获得的时间数据包括绝对时间和相对时间，绝对时间包括几点几分、几月几号等，还可包括农历日期和节假日，例如五月初五、中秋节、父亲节等；相对日期是相对于当前时间的一个偏移量，例如今天是3月1日，根据中间数据中已赋值的字段获得的内容为五天后，则在当前日期往后增加5天，转换后获得的确切时间为3月6日。
本实施例将语音信息的文本信息中可能出现的时间特征放入对应类别的配置文件中，将各类别的配置文件与中间数据中的各字段一一对应，在识别语音信息中的时间信息时，从配置文件中查找匹配的时间特征，并获得匹配时间特征对应的时间符，将该对应的时间符赋值给中间数据中的对应字段，并根据该字段的内容生成确切的时间数据，如此一来，所有待匹配的关键时间特征都存放在各自的配置文件中，识别程序只需要从配置文件中提取时间特征对应的时间符，并通过中间文件转换 获得确切的时间数据，使程序代码的可读性更强，且有效避免了程序代码冗长，在需要修改或添加时间特征数据时，只需要修改配置文件，无须改动和重新编码程序代码，简化了识别程序的修改过程，为程序代码的维护提供了方便。
如图9所示，图9为本申请从语音信息中识别时间信息的装置的第二实施例的模块示意图。本实施例包括图8所示实施例的模块，还包括配置修改模块150，用于：
接收录入的新的时间特征，确定所述新的时间特征对应的类别；
将新的时间特征添加到新的时间特征对应类别的配置文件中；
判断在对应类别的配置文件中是否存在与新的时间特征含义相同的时间特征；
如果是，则根据含义相同的时间特征的时间符设置新的时间特征对应的时间符；
如果否，则对新的时间特征设置新的时间符。
本实施例在修改或增加新的语音识别的关键时间特征时，只需要在对应类别的配置文件中添加新的时间特征，并对新的时间特征设置对应时间符。如果在配置文件中包括有与新的时间特征含义相同的时间特征，则根据含义相同的时间特征的时间符设置新的时间特征对应的时间符；否则，将新的时间符设置为新的时间特征的对应时间符。本实施例在需要修改或添加时间特征数据时，只需要修改配置文件，无须改动和重新编码程序代码，简化了识别程序的修改过程，为程序代码的维护提供了方便。
进一步的，配置查询模块120还用于：
在各个配置文件中的时间特征表中查找与文本信息匹配的至少一个时间特征；
在匹配的时间特征所在的配置文件中的时间符表中查找与匹配的时间特征对应的时间符。
本实施例中的配置文件包括两个表，如图11所示，一个是时间特征表，另一个是时间符表，在建立配置文件时，将该配置文件对应类别的时间特征存储在时间特征表中，而在时间符表中预先设置有多个时间符，将时间特征表中的各个时间特征分别与时间符表中的时间符建立映射关系，时间特征的数量可小于或等于时间符的数量，多个时间特征可对应于一个时间符。在新增时间特征时，将新增的时间特征放置于时间特征表中，并查找该时间特征表中是否有与新增的时间特征含义相同的原有时间特征。如果有，则在时间符表中查找与该含义相同的原有时间特征匹配的时间符，并在查找到的特征符与新增的时间特征之间建立新的映射关系。如果该时间特征表中没有与新增的时间特征含义相同的原有时间特征，则在时间符表中新建一新的时间符，并将新的时间符与新增的时间特征建立映射关系。在查找匹配的时间特征时，在各个配置文件的时间特征表中查找与文本信息匹配的时间特征，然后在该匹配的配置文件的时间符表中查找与匹配的时间特征对应的时间符即可。
由于本实施例的配置文件分别采用了时间特征表和时间符表，将时间特征和时间符分别存储，使匹配查找更加方便。同时，时间特征和时间符的增加分别在各自所在的表中进行，有利于分别扩展时间特征表和时间符表。此外，在多个时间特征对应于一个时间符时，图10所示实施例中需要将同一个时间符在配置文件中重复的出现多次，而图11所示实施例中，无须在配置文件中写入过多的重复数据，只需要在两个表之间建立映射关系即可，无论是时间特征还是时间符，在配置文件中都只出现一次，有利于减少配置文件中的数据量，避免配置文件占用过多存储空间。
进一步的，中间赋值模块130还用于：
当已完成赋值的中间数据中上午/下午字段未赋值，且日期字段未赋值或与当前时间的日期相同时，获取当前时间；
判断已完成赋值的中间数据中时间点字段的时间符是否大于当前时间的时间点；
如果是，则将中间数据的上午/下午字段设置为与当前时间的时段一致；
如果否，则将中间数据的上午/下午字段设置为与当前时间的时段相反。
本实施例在采集的语音信息中，常常会遇到一些不完整的信息，例如“10点提醒我开会”是指的上午十点还是下午十点？这时候则需要优先识别为未来的时间。此时可比较当前时间与语音信息的文本信息的匹配时间特征的大小来判断语音信息中的时段与当前时间的时段是否一致。例如，当前时间为上午9点，当语音信息的文本信息为“10点提醒我开会”时，匹配的时间特征为“10点”，大于当前时间，则默认语音信息中的“10点”为上午10点，提醒闹钟设置为上午10点；如果当语音信息的文本信息为“8点提醒我开会”，匹配时间特征“8点”，小于当前时间，则默认语音信息中的“8点”为下午8点，提醒闹钟设置为下午8点。如此一来，避免了因误解造成时间转换错误，有利于提高语音信息中时间识别的准确性。
进一步的，中间赋值模块130还用于：
当已完成赋值的中间数据中星期字段已赋值，且本周/下周字段未赋值时，获取当前时间；
判断已完成赋值的中间数据中星期字段的时间符是否大于当前时间的星期数；
如果是，则将中间数据的本周/下周字段设置为本周；
如果否，则将中间数据的本周/下周字段设置为下周。
本实施例对应不完整的信息，例如“周三提醒我开会”是指的本周三还是下周三？这时候也需要优先识别为未来的时间。此时可比较当前时间与语音信息的文本信息的匹配时间特征的大小来判断语音信息中的星期数与当前时间的星期数是否一致。例如，当前时间的星期数为周一，当语音信息的文本信息为“周三提醒我开会”时，匹配的时间特征为“周三”，大于当前时间，则默认语音信息中的“周三”为本周的周三，提醒闹钟设置为本周三；如果当前时间的星期数为周五，匹配时间特征“周三”，小于当前时间的星期数，则默认语音信息中的“周三”为下周的周三，提醒闹钟设置为下周三。如此一来，避免了因误解造成时间转换错误，有利于提高语音信息中时间识别的准确性。
进一步的，中间赋值模块130还用于：
当匹配的时间特征中包括明天时，获取当前时间；
在当前时间在上午0点至上午x点范围内时，其中x为预设时间点，则将中间数据中日期字段赋值为与当前时间的日期相同。
本实施例在采集的语音信息中，还会遇到一种特殊的情况，人们往往在凌晨时间段认为是前一天，但实际上凌晨时间段已经是第二天。例如，当前时间是7号凌晨1点，语音信息“明天早上10点提醒我开会”是指的7号上午十点还是8号上午十点？这时候可预先由用户设置一个时间范围，默认该时间范围内为凌晨时间段，在该凌晨时间段内默认为是前一天。例如，预设时间点为5点，则预设的凌晨时间段为上午0点至上午5点，当前时间是7号上午1点(即凌晨1点)，在此范围内，则语音信息“明天早上10点提醒我开会”是指7号上午十点，提醒闹钟设置为7号上午十点。如此一来，避免了因误解造成时间转换错误， 有利于提高语音信息中时间识别的准确性。
需要说明的是，在本文中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者装置不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者装置所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、方法、物品或者装置中还存在另外的相同要素。
上述本申请实施例序号仅仅为了描述，不代表实施例的优劣。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质(如ROM/RAM、磁碟、光盘)中，包括若干指令用以使得一台终端设备(可以是手机，计算机，服务器，或者网络设备等)执行本申请各个实施例所述的方法。
以上所述仅为本申请的优选实施例，并非因此限制本申请的专利范围，凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换，或直接或间接运用在其他相关的技术领域，均同理包括在本申请的专利保护范围内。

Claims (12)

  1. 一种从语音信息中识别时间信息的方法，其特征在于，包括步骤：
    采集语音信息，并识别包含时间信息的语音信息，获得对应于所述语音信息的文本信息；
    识别与所述文本信息匹配的至少一个时间特征，并根据所述匹配的时间特征查找至少一个配置文件以获得对应的时间符；
    将所述对应的时间符赋值给中间数据中与所述匹配的时间特征所在的配置文件对应的字段；所述中间数据包括多个字段，每个字段对应于一个所述配置文件；
    根据所述中间数据中各字段的内容，获得与所述语音信息中的所述时间信息对应的时间数据。
  2. 如权利要求1所述的从语音信息中识别时间信息的方法，其特征在于，进一步包括：
    接收录入的新的时间特征，确定所述新的时间特征对应的类别；
    将所述新的时间特征添加到所述新的时间特征对应类别的配置文件中；
    判断在所述对应类别的配置文件中是否存在与所述新的时间特征含义相同的时间特征；
    如果是，则根据所述含义相同的时间特征的时间符设置所述新的时间特征对应的时间符；
    如果否，则对所述新的时间特征设置新的时间符。
  3. 如权利要求1所述的从语音信息中识别时间信息的方法，其特征在于，所述识别与所述文本信息匹配的至少一个时间特征，并根据所述匹配的时间特征查找至少一个配置文件以获得对应的时间符的步骤包括：
    在各个配置文件中的时间特征表中查找与所述文本信息匹配的至少一个时间特征；
    在所述匹配的时间特征所在的配置文件中的时间符表中查找与所述匹配的时间特征对应的时间符。
  4. 如权利要求1至3任一项所述的从语音信息中识别时间信息的方法，其特征在于，所述根据中间数据中各字段的内容，获得与所述语音信息中的时间信息对应的时间数据的步骤之前还包括：
    当已完成赋值的所述中间数据中上午/下午字段未赋值，且日期字段未赋值或与当前时间的日期相同时，获取当前时间；
    判断已完成赋值的所述中间数据中时间点字段的时间符是否大于当前时间的时间点；
    如果是，则将所述中间数据的上午/下午字段设置为与当前时间的时段一致；
    如果否，则将所述中间数据的上午/下午字段设置为与当前时间的时段不同的时段。
  5. 如权利要求1至3任一项所述的从语音信息中识别时间信息的方法，其特征在于，所述根据中间数据中各字段的内容，获得与所述语音信息中的时间信息对应的时间数据的步骤之前还包括：
    当已完成赋值的所述中间数据中星期字段已赋值，且本周/下周字段未赋值时，获取当前时间；
    判断已完成赋值的所述中间数据中星期字段的时间符是否大于当前时间的星期数；
    如果是，则将所述中间数据的本周/下周字段设置为本周；
    如果否，则将所述中间数据的本周/下周字段设置为下周。
  6. 如权利要求1至3任一项所述的从语音信息中识别时间信息的方法，其特征在于，所述根据中间数据中各字段的内容，获得与所述语 音信息中的时间信息对应的时间数据的步骤之前还包括：
    当所述匹配的时间特征中包括明天时，获取当前时间；
    在当前时间在上午0点至上午x点范围内时，其中x为预设时间点，则将所述中间数据中日期字段赋值为与当前时间的日期相同。
  7. 一种从语音信息中识别时间信息的装置，其特征在于，包括：
    语音采集模块，用于采集语音信息，并识别包含时间信息的语音信息，获得对应于所述语音信息的文本信息；
    配置查询模块，用于识别与所述文本信息匹配的至少一个时间特征，并根据所述匹配的时间特征查找至少一个配置文件以获得对应的时间符；
    中间赋值模块，用于将所述对应的时间符赋值给中间数据中与所述匹配的时间特征所在的配置文件对应的字段；所述中间数据包括多个字段，每个字段对应于一个所述配置文件；
    时间转换模块，用于根据所述中间数据中各字段的内容，获得与所述语音信息中的所述时间信息对应的时间数据。
  8. 如权利要求7所述的从语音信息中识别时间信息的装置，其特征在于，还包括配置修改模块，用于：
    接收录入的新的时间特征，确定所述新的时间特征对应的类别；
    将所述新的时间特征添加到所述新的时间特征对应类别的配置文件中；
    判断在所述对应类别的配置文件中是否存在与所述新的时间特征含义相同的时间特征；
    如果是，则根据所述含义相同的时间特征的时间符设置所述新的时间特征对应的时间符；
    如果否，则对所述新的时间特征设置新的时间符。
  9. 如权利要求7所述的从语音信息中识别时间信息的装置，其特 征在于，所述配置查询模块还用于：
    在各个配置文件中的时间特征表中查找与所述文本信息匹配的至少一个时间特征；
    在所述匹配的时间特征所在的配置文件中的时间符表中查找与所述匹配的时间特征对应的时间符。
  10. 如权利要求7至9任一项所述的从语音信息中识别时间信息的装置，其特征在于，所述中间赋值模块还用于：
    当已完成赋值的所述中间数据中上午/下午字段未赋值，且日期字段未赋值或与当前时间的日期相同时，获取当前时间；
    判断已完成赋值的所述中间数据中时间点字段的时间符是否大于当前时间的时间点；
    如果是，则将所述中间数据的上午/下午字段设置为与当前时间的时段一致；
    如果否，则将所述中间数据的上午/下午字段设置为与当前时间的时段不同的时段。
  11. 如权利要求7至9任一项所述的从语音信息中识别时间信息的装置，其特征在于，所述中间赋值模块还用于：
    当已完成赋值的所述中间数据中星期字段已赋值，且本周/下周字段未赋值时，获取当前时间；
    判断已完成赋值的所述中间数据中星期字段的时间符是否大于当前时间的星期数；
    如果是，则将所述中间数据的本周/下周字段设置为本周；
    如果否，则将所述中间数据的本周/下周字段设置为下周。
  12. 如权利要求7至9任一项所述的从语音信息中识别时间信息的装置，其特征在于，所述中间赋值模块还用于：
    当所述匹配的时间特征中包括明天时，获取当前时间；
    在当前时间在上午0点至上午x点范围内时，其中x为预设时间点，则将所述中间数据中日期字段赋值为与当前时间的日期相同。
PCT/CN2016/086125 2015-07-31 2016-06-17 从语音信息中识别时间信息的方法和装置 Ceased WO2017020657A1 (zh)

Priority Applications (4)

Application Number Priority Date Filing Date Title
EP16832145.3A EP3330962B1 (en) 2015-07-31 2016-06-17 Method and device for identifying time information from voice information
KR1020187000743A KR102046728B1 (ko) 2015-07-31 2016-06-17 음성 정보로부터 시간 정보를 식별하는 방법 및 디바이스
JP2018504856A JP6667931B2 (ja) 2015-07-31 2016-06-17 音声情報から時間情報を認識するための方法およびデバイス
US15/698,423 US10249302B2 (en) 2015-07-31 2017-09-07 Method and device for recognizing time information from voice information

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510465853.7 2015-07-31
CN201510465853.7A CN106409295B (zh) 2015-07-31 2015-07-31 从自然语音信息中识别时间信息的方法和装置

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US15/698,423 Continuation-In-Part US10249302B2 (en) 2015-07-31 2017-09-07 Method and device for recognizing time information from voice information

Publications (1)

Publication Number Publication Date
WO2017020657A1 true WO2017020657A1 (zh) 2017-02-09

Family

ID=57942393

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/086125 Ceased WO2017020657A1 (zh) 2015-07-31 2016-06-17 从语音信息中识别时间信息的方法和装置

Country Status (6)

Country Link
US (1) US10249302B2 (zh)
EP (1) EP3330962B1 (zh)
JP (1) JP6667931B2 (zh)
KR (1) KR102046728B1 (zh)
CN (1) CN106409295B (zh)
WO (1) WO2017020657A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107679823A (zh) * 2017-09-06 2018-02-09 珠海市魅族科技有限公司 一种闹钟提醒方法、装置、计算机装置及计算机可读存储介质
WO2019041343A1 (zh) * 2017-09-04 2019-03-07 深圳传音通讯有限公司 一种闹钟设置方法、用户终端以及电子设备

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107179714B (zh) * 2017-05-17 2020-06-05 广东美的制冷设备有限公司 基于设备控制的时段识别方法和装置
CN107423950B (zh) * 2017-07-07 2021-07-23 北京小米移动软件有限公司 闹钟设置方法和装置
CN108597504A (zh) * 2018-04-04 2018-09-28 上海思依暄机器人科技股份有限公司 一种语音解析方法、装置及机器人
CN108965588A (zh) * 2018-06-27 2018-12-07 维沃移动通信有限公司 一种信息提示方法及移动终端
CN109241271B (zh) * 2018-08-30 2021-09-17 天津做票君机器人科技有限公司 一种汇票交易机器人识别自然语言中汇票到期日的方法
US11727245B2 (en) * 2019-01-15 2023-08-15 Fmr Llc Automated masking of confidential information in unstructured computer text using artificial intelligence
CN110047489A (zh) * 2019-04-04 2019-07-23 科讯嘉联信息技术有限公司 一种家电报装场景智能录入报装时间的方法及系统
CN110619873A (zh) * 2019-08-16 2019-12-27 北京小米移动软件有限公司 音频处理方法、装置及存储介质
WO2024076139A1 (ko) * 2022-10-06 2024-04-11 삼성전자 주식회사 전자 장치 및 상기 전자 장치에서 사용자의 발화를 처리하는 방법
US20240169165A1 (en) * 2022-11-17 2024-05-23 Samsung Electronics Co., Ltd. Automatically Generating Annotated Ground-Truth Corpus for Training NLU Model
CN117725908B (zh) * 2024-02-07 2024-05-24 青岛海尔科技有限公司 一种自然语言处理中的时间解析方法、装置及设备

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101609445A (zh) * 2009-07-16 2009-12-23 复旦大学 基于时间信息的关键子话题提取方法
JP2010218473A (ja) * 2009-03-19 2010-09-30 Casio Computer Co Ltd スケジュール管理装置およびスケジュール管理プログラム
CN102207845A (zh) * 2011-05-25 2011-10-05 盛乐信息技术（上海）有限公司 一种日程表语音管理装置及方法
CN103886093A (zh) * 2014-04-03 2014-06-25 江苏物联网研究发展中心 电子商务搜索引擎同义词的处理方法
CN104021787A (zh) * 2014-06-13 2014-09-03 中国民航信息网络股份有限公司 基于语音识别的机票搜索系统及方法
WO2014168502A1 (en) * 2013-04-10 2014-10-16 Shigabutdinov Ruslan Albertovich Systems and methods for processing input streams of calendar applications

Family Cites Families (23)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0668797B2 (ja) * 1984-05-15 1994-08-31 オムロン株式会社 電子式キヤツシユレジスタ
US6658455B1 (en) * 1999-12-30 2003-12-02 At&T Corp. Method and system for an enhanced network and customer premise equipment personal directory
JP2002023783A (ja) * 2000-07-13 2002-01-25 Fujitsu Ltd 対話処理システム
JP2002215592A (ja) * 2001-01-19 2002-08-02 Just Syst Corp 日時決定装置、日時決定方法、ならびに、記録媒体
US7366979B2 (en) * 2001-03-09 2008-04-29 Copernicus Investments, Llc Method and apparatus for annotating a document
US7246063B2 (en) * 2002-02-15 2007-07-17 Sap Aktiengesellschaft Adapting a user interface for voice control
CN1663249A (zh) * 2002-06-24 2005-08-31 松下电器产业株式会社 元数据生成装置、其生成方法以及检索装置
US20040260561A1 (en) * 2003-01-10 2004-12-23 Joseph Enterprises, Inc. Voice-activated programmable remote control
US20090119062A1 (en) * 2007-11-01 2009-05-07 Timetracking Buddy Llc Time Tracking Methods and Systems
CN101246687A (zh) * 2008-03-20 2008-08-20 北京航空航天大学 一种智能语音交互系统及交互方法
US20100318398A1 (en) * 2009-06-15 2010-12-16 Xerox Corporation Natural language interface for collaborative event scheduling
US8190420B2 (en) * 2009-08-04 2012-05-29 Autonomy Corporation Ltd. Automatic spoken language identification based on phoneme sequence patterns
CN102023995B (zh) * 2009-09-22 2013-01-30 株式会社理光 语音检索设备和语音检索方法
US9922332B2 (en) * 2009-12-09 2018-03-20 Robert Sant'Anselmo Digital signatory and time stamping notary service for documents and objects
US8880406B2 (en) * 2011-03-28 2014-11-04 Epic Systems Corporation Automatic determination of and response to a topic of a conversation
CN102956231B (zh) * 2011-08-23 2014-12-31 上海交通大学 基于半自动校正的语音关键信息记录装置及方法
US8914285B2 (en) * 2012-07-17 2014-12-16 Nice-Systems Ltd Predicting a sales success probability score from a distance vector between speech of a customer and speech of an organization representative
US9172747B2 (en) * 2013-02-25 2015-10-27 Artificial Solutions Iberia SL System and methods for virtual assistant networks
CN104023269A (zh) * 2013-03-01 2014-09-03 联想(北京)有限公司 一种事件提醒方法及电子设备
CN104123937B (zh) * 2013-04-28 2016-02-24 腾讯科技（深圳）有限公司 提醒设置方法、装置和系统
CN103390229A (zh) * 2013-07-23 2013-11-13 宇龙计算机通信科技(深圳)有限公司 一种日程提醒的方法和装置
CN104008476A (zh) * 2014-05-09 2014-08-27 北京奇虎科技有限公司 时间类消息的提醒方法及装置
US20160070580A1 (en) * 2014-09-09 2016-03-10 Microsoft Technology Licensing, Llc Digital personal assistant remote invocation

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2010218473A (ja) * 2009-03-19 2010-09-30 Casio Computer Co Ltd スケジュール管理装置およびスケジュール管理プログラム
CN101609445A (zh) * 2009-07-16 2009-12-23 复旦大学 基于时间信息的关键子话题提取方法
CN102207845A (zh) * 2011-05-25 2011-10-05 盛乐信息技术（上海）有限公司 一种日程表语音管理装置及方法
WO2014168502A1 (en) * 2013-04-10 2014-10-16 Shigabutdinov Ruslan Albertovich Systems and methods for processing input streams of calendar applications
CN103886093A (zh) * 2014-04-03 2014-06-25 江苏物联网研究发展中心 电子商务搜索引擎同义词的处理方法
CN104021787A (zh) * 2014-06-13 2014-09-03 中国民航信息网络股份有限公司 基于语音识别的机票搜索系统及方法

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019041343A1 (zh) * 2017-09-04 2019-03-07 深圳传音通讯有限公司 一种闹钟设置方法、用户终端以及电子设备
CN107679823A (zh) * 2017-09-06 2018-02-09 珠海市魅族科技有限公司 一种闹钟提醒方法、装置、计算机装置及计算机可读存储介质

Also Published As

Publication number Publication date
CN106409295A (zh) 2017-02-15
EP3330962A1 (en) 2018-06-06
JP6667931B2 (ja) 2020-03-18
KR20180016554A (ko) 2018-02-14
JP2018527611A (ja) 2018-09-20
EP3330962A4 (en) 2018-08-08
CN106409295B (zh) 2020-06-16
US20170372702A1 (en) 2017-12-28
US10249302B2 (en) 2019-04-02
KR102046728B1 (ko) 2019-11-19
EP3330962B1 (en) 2021-02-24

Similar Documents

Publication Publication Date Title
WO2017020657A1 (zh) 从语音信息中识别时间信息的方法和装置
US11682380B2 (en) Systems and methods for crowdsourced actions and commands
US20220214775A1 (en) Method for extracting salient dialog usage from live data
CN106782526B (zh) 语音控制方法和装置
US9471666B2 (en) System and method for supporting natural language queries and requests against a user's personal data cloud
CN106020953A (zh) 一种在电子日历中建立日程的方法和装置
US20180366108A1 (en) Crowdsourced training for commands matching
WO2021254046A1 (zh) 一种信息查询方法及装置
WO2011148659A1 (ja) 特別日の登録のための情報処理方法
CN106373572A (zh) 基于人工智能的提示信息的方法及装置
WO2017041377A1 (zh) 用于生成日历提醒信息的方法和装置
CN104202458A (zh) 一种自动保存联系人信息的方法和智能终端
CN103645895B (zh) 自动配置方法以及云编译系统
CN104915345A (zh) 一种推荐服务信息的方法和终端
KR20190107682A (ko) 캘린더 애플리케이션들의 입력 스트림들을 프로세싱하기 위한 시스템들 및 방법들
KR101743731B1 (ko) 분산된 데이터를 통합하여 생성한 온톨로지를 기반으로 쿼리를 처리하는 방법 및 장치
CN105307263B (zh) 个性化定位方法、装置与移动终端
CN117807278A (zh) 基于大语言模型的资源检索方法、训练方法以及装置
CN110716913A (zh) 一种Kafka与Elasticsearch数据库数据的互相迁移方法
CN114172997A (zh) 语音交互方法、装置、电子设备和计算机可读存储介质
CN116088353B (zh) 设备辅助控制方法、装置、设备、存储介质和程序产品
WO2020018826A1 (en) Systems and methods for crowdsourced actions and commands
JP6635460B1 (ja) 情報生成装置、コーパスの生産方法、およびプログラム
WO2015192567A1 (zh) 从文本中提取农历时间的方法、装置及计算机存储介质
WO2022242315A1 (zh) 微件生成方法、装置、电子设备、存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16832145

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 20187000743

Country of ref document: KR

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 2018504856

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE