JP6146703B2 - Voice memo storage method related to schedule - Google Patents

Description

The present invention relates to a voice recording method and a recorded voice registration method in a device having a voice recording function such as an information device having a voice recorder function and a telephone device having a call recording function.

There is a technique for recording a specific part in a conversation by detecting a specific keyword issued during the conversation (for example, Patent Document 1). For example, in a call center, when a voice recognition unit detects a start keyword of a specific product issued by a salesperson during a conversation at a call center, the recording recorder starts recording, and if the customer's understanding is obtained, the salesperson This is an end keyword of the above-mentioned product issued by, which stops the recording of the recorder, and efficiently records and saves only a specific part of the conversation from the start to the end of the conversation.

However, this technology is left as a conversation history in preparation for the occurrence of later troubles caused by conversations related to explanations and questions to customers, such as in-store sales at financial institutions and call centers that accept calls from customers. The recorded data itself is not effectively used for normal work.

JP 2006-107044 A

An object of the present invention is not only to leave a specific part of a conversation but also to an action schedule and a treatment during a conversation in a device having a voice recording function and a function for managing an action schedule and a treatment schedule such as a schedule and a memorandum. To provide a voice memo registration method for a schedule that detects specific terms related to the schedule, and can easily register a partial recording of the conversation related to the action schedule and the treatment schedule promised during the conversation as a voice memo. is there.

In order to solve the above-mentioned problems, the present invention is a method of storing a part of the inputted voice information as a voice memo in a device including voice recording means for recording voice information inputted via a microphone or a network. An activation keyword detection step of analyzing the input voice information and detecting one or more predetermined activation keywords related to the start of the voice memo; and analyzing the input voice information to analyze the date and / or location A date and time location information extraction step for extracting information related to the above, a tail keyword detection step for analyzing the input voice information and detecting one or more predetermined end keywords at the end of the conclusion or message; voice memo storage step for accumulating the voice memo in the information storing means or connected to the device provided in the device If has, in a state where the mode for automatically storing said voice memo is set, when the start keyword detection step detects one of the starting keyword, and starts recording of the audio information, when the trailing keyword detection step detects the silent state of the above case is detected or a certain time one of the last keyword, to stop the recording of the audio information, information before Symbol time location information extraction step according to the date and time and if the extracted information relating to the location, in association with information relating to the most recent date and time and the latest location and the extracted, the accumulated recorded audio information as voice memo relating to planned activity, the extraction step wherein the time location information If the information related to the date and time is not extracted and the information related to the place is not extracted, Characterized by storing a voice memo according to recorded audio information for treatment planning.

According to the present invention, a particular conversation part of the treatment plan and action will be committed during a conversation as a voice memo, as well as automatically extracted, to accumulate classified into the treatment plan and action plans to the voice memo Therefore, there is an advantage that the convenience of a device having a voice recorder function and a call recording function can be improved.

It is a block block diagram of the mobile telephone terminal 1 which concerns on one embodiment of this invention. It is the figure which represented typically the data which the information management part 19 which concerns on one embodiment of this invention manages. It is a flowchart figure showing operation | movement of the audio | voice recording apparatus 1 which concerns on one embodiment of this invention.

Hereinafter, an embodiment of the present invention will be described by taking a mobile phone having a call recording function or a voice recorder function (hereinafter, this apparatus) as an example.

FIG. 1 is a block diagram of the apparatus 1. In FIG. 1, the apparatus 1 includes a terminal control unit 1.
0, speaker 11, microphone 12, operation unit 13, display unit 14, recording / playback control unit 15, recording data storage unit 16, voice recognition processing unit 17, keyword registration unit 18, information management unit 19, and character memo storage unit The memo management unit 190 includes a 191 and a voice memo storage unit 192.

In this embodiment, the character memo accumulating unit 191 and the voice memo accumulating unit 192 are described as being incorporated in the apparatus 1, but may be incorporated in an external device connected to the apparatus 1.

The terminal control unit 10 is a means for controlling the entire apparatus 1 as a mobile phone terminal. The speaker 11 is means for converting the sound information output from the terminal control unit 10 into sound as audible sound and outputting the sound. The microphone 12 is means for converting the voice of the user who uses the apparatus 1 into voice information and outputting the voice information to the terminal control unit 10. The operation unit 13 is means for inputting various key operation information by a user who uses the apparatus 1 to the terminal control unit 10. The display unit 14 is means for presenting various types of information to a user who operates the apparatus 1.

The recording / playback control unit 15 is means for executing recording / playback control related to the voice recorder function or the call recording function provided in the apparatus 1. The recording data storage unit 16 is means for storing the voice information recorded by the recording / playback control unit 15.

The voice recognition processing unit 17 analyzes the voice information from the microphone 12 or the network 2 input via the terminal control unit 10 and detects the activation keyword or the end keyword registered in the keyword registration unit 18 and detects it. The start keyword or the end keyword is notified to the terminal control unit 10.

Further, the voice recognition processing unit 17 analyzes voice information from the microphone 12 or the network 2 input via the terminal control unit 10, extracts date / time information or location information, and uses the extracted activation keyword or end keyword as the terminal Notify the control unit 10. The voice recognition processing unit 1
7 has a built-in dictionary related to date / time information and location information (not shown), and extracts date / time information and location information by a known voice recognition processing technique. Further, the date information and location information extracted by the voice recognition processing unit 17 may be limited to a limited range.

By the way, the voice recognition processing unit 17 may be an unspecified speaker recognition method that does not specify a speaker,
A specific speaker recognition method capable of recognizing only a voice uttered by a specific person using the apparatus 1 may be used.

In particular, in the case of the specific speaker recognition method, it is only necessary to register the speech of words intentionally uttered by a specific person using the device 1, the recognition processing efficiency is high, and the recognition accuracy is also unspecified. There is an advantage that it is higher than the recognition method. In addition, since only the person can use the function of the present application, the apparatus 1
There is also an advantage that an unintended voice memo does not remain even when someone else uses the phone temporarily.
However, since it recognizes only the speech of the words uttered by the specific speaker, it is necessary to repeat and utter even the keyword said by the other party. For example, if the other party is "So then ..."
“I ’ll do it.” I ’ll just repeat it and say the keyword I ’ve registered. This also confirms the promise and does not damage the conversation.

The keyword registration unit 18 uses common words (for example, “Now”, “where”, etc.) used in the beginning part of a schedule or a request, and a common word used in the last part of a schedule or a message. This is means for previously registering words (for example, “waiting”, “please”, etc.) as activation keywords and end keywords.

The information management unit 19 includes a memo management unit 190, a character memo storage unit 191, and a voice memo storage unit 192.
This is a means for registering and managing recording data to be recorded by the recording / playback control unit 15 in accordance with a command input from the terminal control unit 10. In addition, although various schedule memos and miscellaneous memos can be registered in the memo management unit 190, at least an action schedule and a treatment schedule including date and time information can be automatically registered. The automatic registration of the action schedule and the treatment schedule will be described later.

FIG. 2 is a diagram schematically showing the configuration of data managed by the information management unit 19 according to the embodiment of the present invention.

A column 201 is a column for storing the date and time of conversation and information on the other party, and the date and time information is automatically input by a clock function (not shown) built in the terminal control unit 10. In addition, the information of the conversation partner is input automatically or manually. When the conversation is a telephone call via the network 2, the name of the other party is automatically input from the other party's telephone number by a telephone directory function (not shown) built in the terminal control unit 10. Further, the voice recognition processing unit 17 may be provided with a specific speaker identification function so that the conversation partner is automatically determined and input. When inputting manually, the operation part 13 is operated and the name of a partner is selected or operated by key operation.

A column 202 is a column for storing date and time information extracted by the voice recognition processing unit 17 analyzing and extracting voice data during conversation. When voice data including date and time is extracted a plurality of times during conversation, the latest date and time information is valid here.

A column 203 is a column for storing location information extracted by the voice recognition processing unit 17 analyzing and extracting voice data during conversation. When voice data including a location is extracted a plurality of times during a conversation, the latest location information is valid here.

A column 204 is an activation keyword to start recording or an end keyword to stop recording, registered in the keyword registration unit 18, which is extracted by analyzing the voice data during conversation by the voice recognition processing unit 17. This keyword may be a specific word or a special word intentionally issued by a person who uses the apparatus 1. Note that the column 204 is a column for convenience in explaining the present embodiment, and it is not necessary to store it in association with the recorded voice memo, but if it is stored, the voice memo is recorded with any keyword. This is useful when reviewing registered keywords.

Column 205 is a column for storing the contents of voice memos registered in voice memo accumulating unit 192 as voice memos, with the recording data activated by the activation keyword and stopped by the tail keyword.

A column 206 is a type of voice memo, and is a column that stores whether each voice memo in the column 205 is an action schedule, a treatment schedule, or a miscellaneous memo. The action schedule, treatment schedule, and miscellaneous memo are registered in the memo management unit 190. A method for automatically identifying whether the recorded voice memo is an action schedule, a treatment schedule, or a miscellaneous memo will be described later.

Here, the columns 201 to 203 are character data, and the character memo storage unit 1 is used as a character memo.
91 is registered. The column 205 is registered in the voice memo storage unit 192 as a voice memo. The actual voice memo (recorded data) is stored in the recorded data storage unit 16. These character memos and voice memos are associated by the memo management unit 190.

Furthermore, the information managed by the information management unit 19 manages various information other than the scheduled memo (for example, a map related to the action schedule, reference materials related to the treatment schedule, etc.), and each registered content of the memo management unit 190 And various information managed by the information management unit 19 can be associated with each other.

FIG. 3 is an operation flowchart of the apparatus 1 according to the embodiment of the present invention. Less than,
The operation of the apparatus 1 will be described with reference to FIGS. This flow starts with the automatic voice memo mode activated, which automatically extracts voice memos from the voice during conversation (S3).
00).

In the standby state, the terminal control unit 10 determines whether there is a voice input (S301), whether a startup keyword is detected (S310), whether date information is extracted (S320), and whether location information is extracted (S3).
30), the presence / absence of end keyword detection (S340) and the presence / absence of silent detection for a long time (S341) are circulated and monitored.

In the state where the automatic voice memo mode of the apparatus 1 is on, the voice recognition processing unit 17 performs voice information from the microphone 12 or the network 2 or voice information reproduced by the recording / playback control unit 15 via the terminal control unit 10. Input is monitored, and voice information is input (S301,
YES), the input voice information is analyzed, and the presence of the voice information corresponding to the activation keyword is monitored in the voice information (S310).

Here, the voice information input to the voice recognition processing unit 17 is, for example, voice packet data, and the packet size is voice data for a fixed time (data amount variable) or a fixed amount of voice data (recording time). The unit of audio information to be input is arbitrary.

If the activation keyword is detected (S310, YES), the voice recognition processing unit 17 notifies the terminal control unit 10 that the activation keyword has been detected and the detected activation keyword, and the terminal control unit 10 performs the recording / playback control. The unit 15 is controlled to start recording as a voice memo (S311), and the process proceeds to S320.

Although not shown, if YES is obtained in S310, the terminal control unit 10 determines that the information management unit 1
9 and the activation keyword detected by the voice recognition processing unit 17 is stored in the memory of the memo management unit 190 corresponding to the voice memo recorded by the recording / playback control unit 15 (column 204 in FIG. 2).
Corresponding).

In S320, the voice recognition processing unit 17 analyzes the input voice information, monitors the presence of voice information corresponding to the date / time information in the voice information, and extracts the date / time information (S320).
, YES), the terminal control unit 10 is notified that the date / time information has been extracted, and the terminal control unit 10 controls the information management unit 19 to correspond to the voice memo recorded by the recording / playback control unit 15. 1
The extracted date / time information is stored in 90 memories (corresponding to the column 202 in FIG. 2) (S321).
The process proceeds to S330.

In S330, the voice recognition processing unit 17 analyzes the input voice information, monitors the presence of voice information corresponding to the location information in the voice information, and extracts the location information (S330).
, YES), the terminal control unit 10 is notified that the location information has been extracted, and the terminal control unit 10 controls the information management unit 19 to correspond to the voice memo recorded by the recording / playback control unit 15. 1
The extracted location information is stored in 90 memories (corresponding to column 203 in FIG. 2) (S331).
The process proceeds to S340.

In S321 and S331, the date / time information and the location information are stored by overwriting, and the contents stored before that are automatically deleted (overwriting). This
Since the latest date / time information and location information are always stored, the information prior to confirmation extracted in the middle of the conversation is invalid, and the date / time information and location information that was issued last is valid. .

In S340, the speech recognition processing unit 17 analyzes the input speech information, monitors the presence of speech information corresponding to the end keyword in the speech information, and detects the end keyword (YES in S340). The terminal control unit 10 is notified that the end keyword has been detected and the detected end keyword, and the terminal control unit 10 controls the recording / playback control unit 15 to stop recording (S342), and proceeds to S350.

Although not shown, if YES is obtained in S340, the terminal control unit 10 determines that the information management unit 1
9 is stored in the memory of the memo management unit 190 corresponding to the voice memo recorded by the recording / playback control unit 15 (column 204 in FIG. 2).
Corresponding).

In S340, when the speech recognition processing unit 17 does not detect the end keyword (S340)
, NO), the terminal control unit 10 monitors the presence / absence of a predetermined long period of silence (for example, 1 minute), and if the long period of silence is detected (S341, YES), the terminal control unit 10 proceeds to S342 described above. 10 stops the recording of the voice memo (S342), and proceeds to S350. When silence is not detected for a long time (S341, NO), the process returns to S301.

In S350, the terminal control unit 10 stores the date and time information corresponding to S321 and S331.
If there is no date / time information, the voice memo is registered as a miscellaneous memo (S353). If there is date / time information and no location information is stored, the recorded voice memo is scheduled to be treated. (S351), if both date information and place information are stored, the recorded voice memo is registered as an action schedule (S352), and the process proceeds to S360.

It should be noted that here, if date / time information and location information are extracted, it is determined to be an action plan to go to somewhere. Since it is a simple type of automatic identification that is determined as a memo, there is a possibility of erroneous determination, but this registration is provisional registration and can be corrected before the registration is confirmed (described later).

By the way, as the type of voice memo, the setting of whether or not to automatically accumulate each of the action schedule, the treatment schedule, and the miscellaneous memo may be performed by changing the flow so as to pass S351, S352, and S353 (not shown). ).

In S360, the terminal control unit 10 displays information related to the voice memo registered in S351 or S352 (for example, “The next voice memo has been extracted. Do you want to register / edit / delete?”) On the display unit 14. Then, the process proceeds to S361.

In S361, the terminal control unit 10 registers / edits / deletes the information related to the voice memo registered in S351 or S352 according to the operation information input from the operation unit 13, and stores the date / time information stored therein. The storage of the location information is cleared (S362), and the process returns to S301. The registration of the information related to the voice memo is an operation of automatically confirming the temporarily registered contents, and the editing of the information related to the voice memo is performed by editing the text memo associated with the voice memo (columns 201 to 20 in FIG. 2).
3), etc., and association with other information (schedule, memorandum, etc.).

The reason for clearing the storage of the date / time information and the location information is to prevent the date / time information and the location information related to the registered or deleted scheduled memo from being linked to the voice memo extracted later.

The extraction of voice memos according to the above operation flow is continued until the automatic voice memo mode is stopped.

The embodiment of the present invention has been described above. According to the present invention, a specific conversation part related to an action schedule and a treatment schedule promised during a conversation can be automatically extracted as a voice memo and can be easily registered, and the voice memo can be associated with a schedule or a memorandum. Because it is easy
There is an advantage that the convenience of a device having a voice recorder function and a call recording function can be improved.

By the way, although the mobile phone terminal has been described as an example of the embodiment of the present invention, the present invention is not limited to this. The apparatus to which the present invention is applied can be applied to telephone apparatuses in general, such as fixed telephones installed in general offices and homes, and telephone main apparatuses constituting an extension telephone system that accommodates a plurality of extensions. Further, in addition to the telephone device, the present invention can be applied to various information devices such as a personal computer, a personal digital assistant (PDA), a voice recorder, and a smart phone having a voice recorder function.

DESCRIPTION OF SYMBOLS 1 ... Mobile phone terminal 2 ... Network 10 ... Terminal control part 11 ... Speaker 12 ... Microphone 13 ... Operation part 14 ... Display part 15 ... Recording / reproducing control part 16 ... Recording data storage unit 17 ... Voice recognition processing unit 18 ... Keyword registration unit 19 ... Information management unit 190 ... Memo management unit 191 ... Character memo storage unit 192 ... Voice memo Accumulator

Claims (1)

In a device comprising voice recording means for recording voice information inputted via a microphone or a network, a method of storing a part of the inputted voice information as a voice memo,
An activation keyword detection step of analyzing the input voice information and detecting any one or more predetermined activation keywords related to the start of the voice memo , and analyzing the input voice information to determine a date and / or location A date and time location information extracting step for extracting information, a tail keyword detecting step for analyzing the input speech information and detecting one or more predetermined end keywords related to the conclusion or the end of the message, and the speech note anda voice memo storage step of storing the connection information storage unit and the device comprises or said device,
In a state where the mode for automatically storing said voice memo is set, when the start keyword detection step detects one of the starting keyword, and starts recording of the audio information, the tail keyword detection step when detecting the silent state of more or when a predetermined time has detected one of the last keyword, to stop the recording of the audio information,
If before Symbol date and time location information extraction step is to extract the information related to the information and the location according to the date and time, in association with the information relating to the latest date and time and the latest of the location where the extracted, according to the audio information the recording to the action plan If it is stored as a voice memo and the date and time location information extraction step extracts information related to date and time and does not extract information related to location, the recorded voice information is treated in association with the extracted latest date and time information. A voice memo storage method according to a schedule, characterized in that the voice memo is stored as a voice memo according to the schedule.

Images