JPH10198393A - Conversation recording device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a conversation recording device with which recorded results can be easily construed even though conversation among several persons is recorded. SOLUTION: A conversation recording device comprises a voice recognizing means 6, a recognized result storing means 7, a speaker characteristic clustering device 8, a speaker determining means 9 and a display means 10. When the content of conversation is recorded through character data through voice recognition, a time stamp and a voice feature extracted from a voice are also recorded for every character. At the time when the recording is completed, a clustering process is carried out for all voice features in order to obtain a number of persons in the conversation, and typical sound features of speakers. Thus, the voice features of the speakers are compared with recorded data in order to determine a speaker, the content of saying by one and the same speaker is displayed in sections by colors or a displaying position, thereby making it possible carry out a display which can be identified for speakers, respectively.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a conversation recording apparatus for recording conversation contents in a conference or the like.

[0002]

2. Description of the Related Art In recent years, there has been an increasing demand for using a voice recognition device as an input device of a computer or the like, and further application to a conversation recording device for recording the content of a conversation in a conference or the like is expected.

[0003] A conventional conversation recording apparatus will be described below. FIG. 8 is a functional block diagram of a conventional conversation recording device.

In FIG. 8, reference numeral 1 denotes voice input means for converting a voice into an electric signal and inputting the input signal; 2 denotes a voice recognition means for recognizing the electric signal output from the voice input means 1 as a character code; Recognition result storage means 4 for storing the character code output from 2, and display means 4 for displaying the character code stored in the recognition result storage means 4.

[0005] The operation of the conversation recording apparatus configured as described above will be described below. First, the voice during conversation is input as an electrical signal by the voice input unit 1 and sent to the voice recognition unit 2. The voice recognition means 2 compares the waveform of the input electric signal with a human voice and a dictionary storing character codes when the voice is expressed as a character, and recognizes what character was voiced. The results are stored in the recognition result storage means 3 in chronological order, and the contents of the conversation are displayed as characters on the display means 4.

[0006]

However, in such a conversation recording apparatus, when a conversation between a plurality of persons is recorded, a state in which voices of a plurality of persons are mixed is recorded. Had a problem that it was difficult to understand.

That is, as shown in FIG. 9, three speakers, speaker A, speaker B, and speaker C, are, in order, speaker A "this is Z" and speaker B "um". When the user talks with the speaker C, "No, it will be Y", the result of the conversation is displayed on the display means 4 as shown in the screen of FIG. 10, and the conversation between a plurality of persons cannot be identified.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a conversation recording device that allows a user to easily understand the recorded result even when recording conversations between a plurality of persons.

[0009]

SUMMARY OF THE INVENTION In order to achieve this object, the present invention recognizes voice input means for converting voice into an electric signal and inputs the same, and recognizes the electric signal output from the voice input means as a character code. Voice recognition means, a character code output from the voice recognition means, a voice feature used for recognition, and a recognition result storage means for storing the recognition time in pairs, and all of the recognition results stored in the recognition result storage means. Among the features of speech, speech features having similar features are grouped together, and speaker feature clustering means for obtaining the features of the speakers present during the conversation, and speaker features obtained by the speaker feature clustering means are obtained. Speaker discriminating means for discriminating a speaker who has uttered a character code paired with the speech feature by comparing with a feature of the speech stored in the recognition result storage means, Person based on the determination result obtained by the determining means, the same speaker is configured to have a display means for displaying a character so as to be visually identify the character code of the recognition result of the speech uttered.

As a result, even when a conversation between a plurality of persons is recorded, a conversation recording device in which the recorded result can be easily understood can be obtained.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS According to the first aspect of the present invention, there is provided a voice input means for converting a voice into an electric signal and inputting the same, and a voice for recognizing the electric signal output from the voice input means as a character code. Recognition means, recognition result storage means for storing the character code output from the voice recognition means, the voice feature used for recognition, and the recognition time in pairs, and all the voices stored in the recognition result storage means Speaker characteristics clustering means for obtaining the characteristics of speakers present during a conversation, and speaker characteristics obtained by the speaker characteristics clustering means. A speaker discriminating means for discriminating a speaker who has uttered a character code paired with the speech feature by comparing with a feature of the speech stored in the recognition result storage means; Display means for displaying the character code of the recognition result of the voice uttered by the same speaker as characters based on the determination result obtained in The content of the conversation of the speaker can be recorded, and the content of the conversation of the same speaker can be displayed for each speaker.

According to a second aspect of the present invention, there is provided a voice input means for converting a voice into an electrical signal and inputting the voice signal, a voice recognition means for recognizing the electrical signal output from the voice input means as a character code, Character code output from the voice recognition means, the characteristics of the voice used for recognition, and the recognized time,
Recognition result storage means that is stored as a pair, and speech features having approximate features among all the speech features stored in the recognition result storage means are combined to obtain the features of the speaker present during the conversation. A character paired with the speech feature by comparing the speaker feature clustering means and the speech feature stored in the recognition result storage means with reference to the speaker feature obtained by the speaker feature clustering means. Speaker discriminating means for discriminating a speaker who uttered the code;
A speaker selecting means for selecting and outputting only a character code of a recognition result of a voice uttered by the same speaker based on a discrimination result obtained by the speaker discriminating means, and a character output from the speaker selecting means With a configuration having display means for displaying codes as characters, it is possible to record the contents of conversations of a plurality of speakers and display only the conversation contents of the speaker selected by the user on the display means.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. (Embodiment 1) FIG. 1 is a functional block diagram of a conversation recording device according to a first embodiment of the present invention, and FIG. 2 is a circuit block diagram of the conversation recording device according to the first embodiment of the present invention.

In FIG. 1, reference numeral 5 denotes voice input means for converting a voice into an electric signal and inputting it; 6, a voice recognition means for recognizing the electric signal output from the voice input means 5 as a character code; Character code output from 6,
Recognition result storage means for storing the features of the voice used for recognition and the time of recognition in pairs, respectively, and among all the voice features stored in the recognition result storage means 7, a voice having an approximate feature is stored. A speaker characteristic clustering means 9 for obtaining the characteristics of the speakers present in the conversation collectively, and a speech stored in the recognition result storage means 7 based on the speaker characteristics obtained by the speaker characteristic clustering means 8. The speaker discriminating means for discriminating the speaker who uttered the character code paired with the voice feature in comparison with the feature, and 10 based on the discrimination result obtained by the speaker discriminating means 9, This is display means for displaying a character code of a recognition result of a voice uttered by a person as a character so that the character code can be visually identified.

In FIG. 2, reference numeral 12 denotes a microphone for converting collected voice into an electric signal, and reference numeral 13 denotes an analog-digital converter for converting an analog electric signal obtained from the microphone 12 into a digital signal.

Reference numeral 14 denotes a central processing unit (hereinafter referred to as "CP") for executing a control program stored in a ROM 15 described later.
U), 15 is a read-only memory (hereinafter referred to as “ROM”) for fixedly storing data, and is a control program 1 executed by the CPU 14.
6 and speech recognition dictionary data 17 are stored. Reference numeral 18 denotes a random access memory (hereinafter, referred to as “RAM”) for temporarily storing data, and an area for storing recognition result storage data 19 and speaker discrimination result storage data 20. Have.

Reference numeral 21 denotes a keyboard for externally giving commands such as start and end to the CPU 14, and 22 denotes a display device for displaying a conversation record result, such as a CRT display or an LCD (liquid crystal display). Having. 23 is a bus line for signal transmission inside the device.

The operation of the conversation recording apparatus configured as described above will be described below with reference to the flowchart of FIG.

FIG. 3 is a flowchart of the conversation recording device according to the first embodiment of the present invention. In the present embodiment, as shown in FIG. 9, three speakers A, B, and C are, in order, speaker A “this is Z” and speaker B “er”. The case where the user is talking with the speaker C, "No, Y will" will be described as an example.

As shown in FIG. 3, voice data obtained by taking in conversations by a plurality of speakers by the microphone 12 is converted into a digital voice signal by the analog-digital converter 13 and stored in the RAM 18 (S1). In the case of FIG. 9, an audio digital signal of "This is Z, is it Y?" Is recorded in the RAM 18. At the time of voice data capture, it is also possible to specify the number of speakers.

Next, by applying a calculation formula to the digital voice signal stored in the RAM 18 in advance, voice breaks, voice characteristics, and the like are extracted (S2).

The speech feature extracted from the waveform of the input electric signal is compared with a human utterance and a dictionary storing character codes when the utterance is expressed as a character to recognize what character is uttered. (S3).

At the time of voice recognition, the character code obtained by the voice recognition, the characteristics of the voice used for the recognition, and the recognition time are stored as a pair in the RAM 18 as recognition result storage data 19 (S4).

Further, of all the voice features stored in the RAM 18, voice features having approximate features are grouped together, and the features of the speaker present during the conversation are obtained (S).
5). If the number of speakers is input at the time of capturing the voice data in S1, it is possible to improve the accuracy of the clustering of the voice feature using the number information of the speakers.

Based on the speaker characteristic obtained by the speaker characteristic clustering, the speaker which has uttered the character code paired with the voice characteristic is discriminated by comparing it with the voice characteristic stored in the RAM 18. (S6).

Based on the discrimination result, the character code of the recognition result of the voice uttered by the same speaker is displayed as a character so that it can be visually identified (S7), and all the processing ends. FIG. 4 shows this display state, and shows the result displayed for each speaker.

(Embodiment 2) FIG. 5 is a functional block diagram of a conversation recording apparatus according to a second embodiment of the present invention.

In FIG. 5, reference numeral 11 denotes a speaker selecting means for selecting a character code of the recognition result of the voice uttered by the speaker specified by the user from the result of the determination obtained by the speaker determining means 9. Other means have been described in the first embodiment (see FIG. 1), and thus description thereof will be omitted.

The circuit configuration is the same as that of the first embodiment (see FIG. 2).
This is realized by the user operating the keyboard 21.

The operation of the conversation recording device configured as described above will be described below with reference to the flowchart of FIG.

FIG. 6 is a flowchart of the conversation recording device according to the second embodiment of the present invention.

As shown in FIG. 6, voice data obtained by taking in conversations by a plurality of speakers by the microphone 12 is converted into a digital voice signal by the analog-digital converter 13 and stored in the RAM 18 (S8). In addition,
At the time of voice data capture, it is also possible to specify the number of speakers.

Next, a predetermined formula is applied to the digital voice signal stored in the RAM 18 to extract voice breaks, voice features, etc. (S
9).

The speech feature extracted from the waveform of the input electric signal is compared with the utterance of a person and a dictionary storing character codes when the utterance is expressed as a character, and it is recognized what character was uttered. (S10).

At the time of speech recognition, the character code obtained by the speech recognition, the characteristics of the speech used for the recognition, and the recognition time are stored as a pair in the RAM 18 as recognition result storage data 19 (S11).

Further, of all the voice features stored in the RAM 18, voice features having approximate features are grouped together, and the features of the speaker present during the conversation are obtained (S1).
2). If the number of speakers is input at the time of capturing the voice data in S1, it is possible to improve the accuracy of the clustering of the voice feature using the number information of the speakers.

Based on the speaker characteristics obtained by the speaker characteristic clustering, the speaker that has uttered the character code paired with the voice characteristics is discriminated by comparing with the voice characteristics stored in the RAM 18. (S13).

Here, a message urging the designation of the speaker to be displayed is displayed on the display device 22, and a key input from the user is waited (S14).

If there is a key input, the process proceeds to S15. In S15, it is determined which speaker has been designated by the user. If no speaker is specified, the process proceeds to S17, and if a speaker is specified, the process proceeds to S16.

In S16, only the character code of the recognition result of the voice uttered by the designated speaker is displayed as characters so that it can be visually identified. FIG. 7 shows an example in which speaker B is specified.

If the speaker is not specified in S15, the character code of the recognition result of the voice uttered by the same speaker for each speaker is displayed on the display device 22 as characters so that it can be visually identified (S17). ).

[0042]

As described above, according to the present invention, the voice data obtained by the voice input means is converted into a character code by the voice recognition means, and further, based on the voice characteristics of the speaker,
By providing a means for classifying the character codes obtained by the voice recognition means for each speaker,
An advantageous effect is obtained in that it is possible to display the image separately by color or display position.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of a conversation recording device according to a first embodiment of the present invention.

FIG. 2 is a circuit block diagram of a conversation recording device according to the first embodiment of the present invention.

FIG. 3 is a flowchart of a conversation recording device according to the first embodiment of the present invention.

FIG. 4 is an explanatory diagram of a screen display of the conversation recording device according to the first embodiment of the present invention.

FIG. 5 is a functional block diagram of a conversation recording device according to a second embodiment of the present invention.

FIG. 6 is a flowchart of a conversation recording device according to a second embodiment of the present invention.

FIG. 7 is an explanatory diagram of a screen display of a conversation recording device according to a second embodiment of the present invention.

FIG. 8 is a functional block diagram of a conversation recording device in a conventional example.

FIG. 9 is a diagram showing a state of a conversation in a conventional example.

FIG. 10 is an explanatory diagram of a screen display of a conversation recording device in a conventional example.

[Explanation of symbols]

 Reference Signs List 5 voice input means 6 voice recognition means 7 recognition result storage means 8 speaker feature clustering means 9 speaker determination means 10 display means 11 speaker selection means 12 microphone 13 analog-digital conversion device 14 central processing unit (CPU) 15 read-only Memory (ROM) 16 Control program 17 Voice recognition data 18 Random access memory (RAM) 19 Recognition result storage data 20 Speaker discrimination result storage data 21 Keyboard 22 Display device 23 Bus line

Claims (2)

[Claims] A voice input unit for converting a voice into an electric signal and inputting the voice signal; a voice recognition unit for recognizing the electric signal output from the voice input unit as a character code; and a character output from the voice recognition unit A recognition result storage unit that stores the code, the voice feature used for recognition, and the recognition time in pairs, and a voice having an approximate feature among all voice features stored in the recognition result storage unit. A speaker feature clustering unit that collects the features and obtains the characteristics of the speakers present during the conversation, and is stored in the recognition result storage unit based on the speaker characteristics obtained by the speaker characteristic clustering unit. A speaker discriminating means for discriminating a speaker who has uttered the character code paired with the speech feature by comparing with the speech feature. Determination Based on the results, the conversation recording apparatus; and a display unit for the same speaker is displayed as a character so that it can be visually identify the character code of the recognition result of the speech uttered. 2. A voice input means for converting a voice into an electric signal for input, a voice recognition means for recognizing an electric signal output from the voice input means as a character code, and a character output from the voice recognition means. A recognition result storing means for storing the code, the feature of the voice used for recognition, and the time of recognition in pairs, and having an approximate feature among all voice features stored in the recognition result storing means. A speaker feature clustering unit that collects voice features and obtains features of a speaker present during a conversation, and is stored in the recognition result storage unit based on the speaker features obtained by the speaker feature clustering unit. Speaker discriminating means for discriminating a speaker who has uttered a character code paired with the speech feature by comparing the obtained speech feature with the obtained speech feature. Speaker selecting means for selecting and outputting only the character code of the recognition result of the voice uttered by the same speaker based on the determination result, and displaying means for displaying the character code output from the speaker selecting means as characters And a conversation recording device.