KR101687614B1 - Method for voice recognition and image display device thereof - Google Patents

Abstract

The present invention relates to a voice recognition method and a video display apparatus according to the present invention in which a user can more easily control an image display apparatus by increasing the voice recognition rate.
The speech recognition method and the image display apparatus according to an embodiment of the present invention can increase the voice recognition rate. Accordingly, the image display apparatus can be remotely controlled more accurately.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to a voice recognition method and an image display device,

The present invention relates to a speech recognition method and a video display apparatus therefor.

And more particularly, to a speech recognition method having a high voice recognition rate and a video display device therefor.

The video display device is a device capable of displaying a predetermined video, and includes a digital television, a set-top box, a PVR (Personal Video Recorder), or a DVD video player. A remote control device or the like can be used for the user to operate such a video display device from a long distance.

Beyond conventional analog broadcasting, digital-based digital broadcasting technology is being developed and commercialized. Accordingly, it is possible to provide various types of contents services to users, such as real-time broadcasting, contents on demand (CoD), games, or news, using an Internet network connected to each home in addition to existing radio waves or cable cable media. In addition, the digital broadcast receiver can receive and display various contents services described above. An example of content service provision using the Internet network is IPTV (Internet Protocol TV).

As described above, due to the development of the digital broadcasting technology and the provision of the IPTV service, the video display device can be provided with a wide variety of services, and games and the like can be used through the video display device. As the contents or services that can be provided by the video display device are greatly diversified, the number of keys of a remote controller necessary for using a predetermined content or service is also increasing. Each of the keys has various and complicated control commands .

Accordingly, a voice recognition remote controller capable of inputting a control command by voice recognition in addition to a button-type remote controller including only limited keys has been developed.

1 is a diagram for explaining control of a video display device by voice recognition.

Referring to FIG. 1, a user inputs a voice signal including a predetermined command through a remote controller 120 having a voice recognition function. Then, the remote controller 120 transmits the voice signal to the video display device 110, and the video display device 110 recognizes the command included in the voice signal and performs the control operation accordingly.

2 is a flowchart for explaining a general operation of speech recognition.

Referring to FIG. 2, in a general speech recognition method, a voice signal is first input to a remote controller 120 (step 210).

The video display device 110 receiving the input voice signal discriminates words included in the voice signal (step 220). That is, it recognizes a command included in the voice signal.

Then, a control operation is performed according to the word recognition result (operation 230).

A remote controller for remotely controlling an image display apparatus by voice recognition can flexibly input various control keys, but if an error occurs in recognition of a voice signal, a wrong control key can be input.

Accordingly, there is a need to provide a speech recognition method capable of speech recognition with high speech recognition rate and high accuracy, and a video display device accordingly.

The present invention aims to provide a speech recognition method capable of increasing the speech recognition rate and a video display device therefor.

It is another object of the present invention to provide a speech recognition method and a video display apparatus therefor, in which a user can more easily control a video display device by increasing the voice recognition rate.

According to an embodiment of the present invention, there is provided a speech recognition method comprising: receiving a speech signal corresponding to a one-shot type character at predetermined time intervals; Taking a mouth-shaped motion image; Extracting a voice signal section corresponding to the one-letter type character from the voice signal, and extracting a video signal section corresponding to the one-letter type character from the motion picture; And recognizing the completion type character corresponding to at least one of the voice signal and the video signal input in the matched section if the voice signal section and the video signal section coincide with each other.

The recognizing of the completion type character may include calculating a speech characteristic value of the speech signal input in the matched section; And recognizing the complete character based on the voice feature value.

The step of recognizing the completion character may further include calculating a motion feature value of the motion image input in the matched section.

In addition, the step of recognizing the complete character based on the voice feature value

And recognizing the complete character based on the voice feature value and the motion feature value.

The step of receiving the voice signal may include inputting at least one of the voice signals corresponding to at least one completion character forming the predetermined command at the predetermined time intervals.

The step of extracting the video signal section may include extracting a section from the time when the audio signal starts to be input until the beginning of the predetermined time interval into the audio signal section. And extracting a section from the time point at which the mouth shape starts moving to the moment at which the mouth shape is stopped into the video signal section.

According to another aspect of the present invention, there is provided a speech recognition method including: outputting the predetermined command information including the recognized complete character or the completed character through a user interface in real time; And performing the predetermined command.

In addition, the step of recognizing the completion type character may include transmitting the voice signal and the video signal input in the matched section to the Internet server. And recognizing the completion type character using the speech recognition engine and the speech recognition database of the Internet server.

According to another embodiment of the present invention, there is provided a method of recognizing speech, comprising the steps of: outputting the speech similarity through a user interface when at least one speech similarity exists in the recognized completed character; And receiving a completion character intended by a user of the at least one voice similarity through the user interface.

According to another aspect of the present invention, there is provided a speech recognition method comprising: outputting a plurality of predetermined commands through a user interface when a plurality of predetermined commands are recognized; And receiving a predetermined command from the plurality of predetermined commands through the user interface.

According to another aspect of the present invention, there is provided a speech recognition method including: outputting the predetermined command information including the recognized complete character or the completed character through a user interface in real time; And searching the Internet server for information related to the predetermined command when the input of the voice signal is completed, and outputting the retrieved information through the user interface.

The image display apparatus according to an embodiment of the present invention includes a remote controller for receiving a voice signal corresponding to a one-shot type character at a predetermined time interval and photographing a mouth-shaped motion image; Extracting a voice signal section corresponding to the one-letter type character from the voice signal, extracting a video signal section corresponding to the one-letter type character from the motion image, and, if the voice signal section and the video signal section coincide, A voice recognition processor for recognizing the completion type character corresponding to at least one of the voice signal and the video signal inputted within a predetermined interval; And a control unit for controlling the execution of the predetermined command including the recognized complete character.

The speech recognition processing unit may calculate a speech characteristic value of the speech signal input in the matched section and recognize the one-sided character based on the calculated speech characteristic value.

The speech recognition method and the image display apparatus according to an embodiment of the present invention can increase the voice recognition rate. Accordingly, the image display apparatus can be remotely controlled more accurately.

In addition, the speech recognition method and the image display apparatus according to an embodiment of the present invention can increase the voice recognition rate, so that the user can more easily and conveniently use the image display apparatus.

1 is a diagram for explaining control of a video display device by voice recognition.
2 is a flowchart for explaining a general operation of speech recognition.
3 is a block diagram illustrating an image display apparatus according to an exemplary embodiment of the present invention.
FIG. 4 is a block diagram illustrating FIG. 3 in more detail.
5 is a view illustrating a display screen output by the image display apparatus according to an embodiment of the present invention.
6 is a view illustrating another display screen output by the image display apparatus according to an embodiment of the present invention.
7 is a view illustrating another display screen output by the image display apparatus according to the embodiment of the present invention.
8 is a view illustrating another display screen output by the image display apparatus according to an embodiment of the present invention.
9 is a view showing another display screen output by the image display apparatus according to an embodiment of the present invention.
10 is a view illustrating another display screen output by the image display apparatus according to an embodiment of the present invention.
11 is a diagram illustrating a speech recognition method according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings and accompanying drawings, but the present invention is not limited to or limited by the embodiments.

As used herein, terms used in the present invention are selected from general terms that are widely used in the present invention while taking into account the functions of the present invention, but these may vary depending on the intention or custom of a person skilled in the art or the emergence of new technologies. Also, in certain cases, there may be a term chosen arbitrarily by the applicant, in which case the meaning thereof will be described in the description of the corresponding invention. Therefore, it is intended that the terminology used herein should be interpreted based on the meaning of the term rather than on the name of the term, and on the entire contents of the specification.

3 is a block diagram illustrating an image display apparatus according to an exemplary embodiment of the present invention.

3, an image display apparatus 300 according to an exemplary embodiment of the present invention includes a signal input unit 310, a voice recognition unit 320, a controller 330, a signal processing unit 340, An OSD generation unit 355, a user interface unit 360, and a storage unit 380. The OSD generation unit 355, the OSD generation unit 355, the user interface unit 360, In addition, it further includes a remote controller unit 390 externally.

The video display device 300 is a digital broadcast receiver such as a digital television or a set-top box. When the image display apparatus 300 is a digital television, the image display apparatus 300 includes a display unit 370 as shown in FIG. Also, when the image display apparatus 300 is a set-top box, the image display apparatus 300 does not include a display unit 370.

In addition, the image display apparatus 300 may further include other configurations required in addition to the configuration shown in FIG.

The signal input unit 310 may include a tuner 311, a network interface unit 315, and the like. Hereinafter, a signal received by the signal input unit 310 is referred to as a video signal.

The tuner 311 selectively receives a broadcast signal transmitted in the form of a radio frequency (RF) signal through a channel of a predetermined frequency band. That is, it selectively receives a broadcast signal transmitted from a content producer such as a broadcasting station and including a predetermined content.

The network interface unit 315 accesses a predetermined Internet server or a content provider (CP) server through a network, and transmits / receives a predetermined signal to / from the servers. Specifically, the network interface unit 315 receives a broadcast signal including a predetermined content from a content provider.

The remote control unit 390 receives a voice signal corresponding to a one-time type character at a predetermined time interval, and captures a mouth-shaped motion image. Then, the input voice signal and the captured motion image are transmitted to the interface unit 350. The remote control unit 390 may include a camera unit 391, a microphone unit 393, a signal conversion unit 395, and a remote control interface unit 397.

One complete character is a character that can be used independently, and there are 2350 characters (2850 of extended type completed characters) in the Hangul complete character, and 26 characters There is a character. Here, the Hangul complete character can be composed of at least one consonant (basic consonant and compound consonant) and at least one vowel (basic vowel and complex vowel).

And each completed character has different pronunciation characteristics. For example, the 'H' and 'I' characters have distinct pronunciation characteristics of 'I' and 'I', respectively, and 'a' and 'b' Has a pronunciation characteristic of 'bi'.

 In the following description, the completion type character is used as an example of the completion type character in Korean or English, but the completion type character according to another foreign language can be used.

The camera unit 391 takes a mouth-shaped image when the user speaks a predetermined command. That is, the user takes a mouth-shaped image when the user inputs a voice signal to the remote control unit 390. The image captured by the camera unit 391 may be a moving picture.

The microphone unit 393 receives a voice signal generated when the user speaks a predetermined command. That is, the microphone unit 393 records a voice signal generated by the user.

The signal converting unit 395 converts the video signal captured by the camera unit 391 into a predetermined signal format necessary for transmission to the interface unit 350. In order to transmit the voice signal input from the microphone unit 393 to the interface unit 350, the voice signal is converted into a predetermined signal form.

For example, if the remote control unit 390 and the interface unit 350 include a radio frequency (RF) module capable of transmitting and receiving a predetermined signal according to the RF communication standard, the signal conversion unit 395 converts the video signal or the audio signal Into an RF signal conforming to the RF communication standard. Alternatively, if the remote control unit 390 and the interface unit 350 include an IR module capable of transmitting and receiving a predetermined signal in accordance with the infra-red communication standard, the signal conversion unit 395 converts the video signal or the audio signal Converts to IR signal according to IR communication standard.

The remote controller interface unit 397 transmits the signal output from the signal converter 395 to the interface unit 350. The remote control interface unit 397 transmits and receives signals according to an RF (radio frequency) module (not shown) or an IR (infra-red) communication standard that can transmit and receive signals according to the RF communication standard And an IR module (not shown).

The interface unit 350 receives the voice signal and the video signal transmitted from the remote controller interface unit 397 and transmits the voice signal and the video signal to the voice recognition processor 320. The interface unit 350 may include an IR module capable of transmitting and receiving signals according to an RF communication standard or an IR module capable of transmitting and receiving signals according to an IR (infra-red) communication standard (not shown) Time).

The voice recognition processor 320 extracts a voice signal section corresponding to the one-letter type character from the received voice signal and extracts a video signal section corresponding to the one-letter type character from the mouth-shaped motion image. If the extracted voice signal section and the video signal section coincide with each other, the completion type character corresponding to at least one of the voice signal and the video signal input in the matched section is recognized.

Detailed operation of the speech recognition processor 320 will be described in detail with reference to FIG.

The control unit 330 controls the overall operation of the video display device 300. Specifically, the video signal including the predetermined content is controlled to be displayed on the video screen. Specifically, the control unit 330 controls the voice recognition processor 320 to execute a predetermined command including the completed character.

The signal processor 340 converts the received video signal into data that can be displayed by the video display device 300 and outputs the data. Specifically, it demodulates the received video signal, demultiplexes and decodes the demodulated signal, and performs signal processing such as error correction and signal quality improvement.

The OSD generation unit 355 generates OSD (On Screen Data) data and outputs it to the display unit 370. Specifically, the OSD generation unit 355 can convert the user interface data output from the user interface unit 360 into OSD data and output the OSD data. Also, according to the control of the control unit 330, various information to be provided to the user can be generated as the OSD data. Specifically, the recognition information of the completion type character corresponding to the voice signal input to the user is generated and output in real time on the OSD. The generated OSD is transmitted to the display unit 370 and displayed on the display screen.

The user interface unit 360 generates and outputs a control menu to be provided to the user as user interface (UI) data, or receives a predetermined request or predetermined information from the user. Here, the user interface data may be converted into OSD (On Screen Data) data through the OSD generating unit 355. Also, the user interface data may be directly transmitted to the display unit 370 and displayed in a GUI (Graphic User Interface).

The OSD data output from the OSD generation unit 355 or the user interface UI output through the user interface unit 360 will be described in detail with reference to FIG. 5 to FIG.

The display unit 370 displays a video signal transmitted from the signal processing unit 340 as an image screen. In addition, the OSD data or the user interface data output from the user interface unit 360 or the OSD generating unit 355 is displayed on the whole or a predetermined area of the image screen.

The storage unit 380 may store various kinds of information necessary for the display operation under the control of the controller 330. [ The storage unit 380 may store a speech recognition database for speech recognition.

The image recognition operation and the detailed configuration of the image display apparatus 300 according to an embodiment of the present invention will be described in more detail with reference to FIG. 4 to FIG.

FIG. 4 is a block diagram illustrating FIG. 3 in more detail.

4, the speech recognition processing unit 320 may include a voice and video signal input unit 420, a section detection unit 430, a feature value calculation unit 440, and a character recognition unit 450. In addition, the network interface unit 315 can transmit / receive predetermined data to / from the Internet server 410 via a wired or wireless communication network. Here, the Internet server 410 includes a recognition database 411 and a recognition engine 413. Description of the overlapping configuration in FIG. 3 is omitted.

The interface unit 350 receives the audio signal and the mouth-shaped motion picture signal transmitted from the remote control unit 390. Hereinafter, a mouth-shaped motion picture signal is referred to as a 'video signal'.

The audio and video signal input unit 420 receives the audio and video signals input from the remote control unit 390 from the interface unit 350.

The remote control unit 390 receives a voice signal from the microphone unit 393, and receives a voice signal corresponding to a one-time type character at predetermined intervals. For example, when a voice signal of 'channel 7' is input, 'ch` is inputted,' null 'is input after a predetermined time, and' 7 (seven) 'is input after a predetermined time.

Here, the predetermined time means a stop time that exists between one complete character and another successive character that is subsequently input. In other words, in inputting the completion type characters (Chain, 7, 7) in the reading format, the user's voice signal is not input at the stopping time (predetermined time interval). Accordingly, it can be determined that the section in which the signal in the frequency domain corresponding to the human voice is not input is the above-described stop time.

Hereinafter, 'one-time completion type character + predetermined time interval + one-time completion character time + predetermined time interval ...' is expressed as 'one-time completion type character completion character type _...'. That is, the predetermined time interval is displayed using the symbol '_'. For example, 'chess', 'null', and 'chess', which are input at predetermined time intervals, can be expressed as 'chessboard (7)'.

In the video signal photographed by the camera unit 391, there is a section in which the mouth shape does not move in a mouth shape of a user who speaks one complete character and a succeeding complete character. The section where the mouth shape does not move is the predetermined time interval described above.

The section detecting unit 430 extracts a voice section and an image section corresponding to the one-letter type character. The section detector 430 may include a voice section detector 431 and an image section detector 433. [

The voice section detector 431 extracts a voice signal section corresponding to the one-letter type character from the voice signal. Specifically, a section between adjacent stopping times can be extracted as a speech signal section corresponding to a one-time-type character. For example, if a voice signal starts to be input at time point a and a stop time is detected at time point b, a section from time point a to time point b can be extracted as a voice signal section corresponding to a one-time type character. If the stop time is detected from the time point b to the time point c and the voice signal is detected again from the time point c to the time point d, the voice signal section corresponding to the succeeding one- do.

Then, the video segment detecting unit 433 detects an video segment corresponding to the one-shot type character from the video signal. Specifically, a section between adjacent still image sections can be extracted as a video signal section corresponding to a one-shot type character. For example, if a mouth shape starts to move from a point of time and a still image portion in which a mouth motion is stopped from point b is detected, a section from a point of time to point b is referred to as a video signal section corresponding to a full- Can be extracted.

In an image display apparatus according to an embodiment of the present invention, a voice signal is input in a 'full character type + predetermined time interval', and a voice signal section and a video signal section are detected. Since the voice signal is recognized in a section where the voice signal section and the video signal section coincide with each other, the voice signal can be recognized more accurately. Specifically, even if the audio signal is input uncertainly due to the ambient noise, the audio signal can be received more accurately by comparing the duration of the video signal and detecting the duration of the audio signal that coincides with the video signal duration. Accordingly, the recognition rate of the voice signal can be increased.

If the voice signal section corresponding to the one-time type character detected by the voice section detecting section 431 coincides with the video signal section corresponding to the one-time type character detected by the video section detecting section 433, the feature value calculating section 440 calculates The feature value of the voice signal existing within the matching section detected by the detecting section 430 is calculated. Further, the feature value of the video signal can be further calculated. Hereinafter, the audio signal input section when the audio signal section and the video signal section match is referred to as a " matching section ".

More specifically, the voice characteristic value calculated by the voice characteristic calculating unit 441 is a value obtained by digitally converting the voice signal existing in the matching period and calculating a Mel Frequency Cepstral Coefficient (MFCC) value obtained by a discrete cosine transform (DCT) And the like. The voice characteristic value is not the voice signal itself but only a characteristic value extracted from the voice signal and has a smaller data amount (for example, 10% of the original voice signal data amount) than the original voice signal.

Specifically, the voice characteristic calculator 441 receives a voice signal in the form of an analog signal and converts it into a PCM (Pulse Code Modulation) signal using an analog-to-digital converter (AD converter). The converted PCM signal has a sampling rate of 8 KHz and can have an amplitude resolution of 16 bits. Subsequently, a noise component is removed from the PCM signal, leaving only a frequency component of a section corresponding to a human voice. Then, the PCM signal is divided for each predetermined frequency band, and the MFCC value in each divided frequency band is obtained. The feature values of the speech signal can be obtained by various other methods as well.

Further, the motion feature calculating unit 443 extracts the lip motion, the jaw motion, and the cheek motion from the video signal existing in the matching period, and calculates the feature value of each motion direction and the movement amount. Specifically, the motion of the lips, the jaws, or the cheeks can be calculated as a motion vector value.

A discrete signal obtained by frequency conversion of a voice signal itself is used, the size of a voice signal itself becomes very large, and a time and a data transmission amount for completing a transmission to a predetermined Internet server through the network are greatly increased . Accordingly, the time of speech recognition is increased, and accordingly, it is difficult to control according to immediate speech recognition.

In this embodiment, only the feature value of the speech signal is calculated by the speech feature calculating unit 441, and only the feature value of the motion is calculated by the motion feature calculating unit 443. Since the extracted feature value is transmitted to the Internet server 410 through the network interface unit 310, the time and data transfer time required to complete the transmission to the Internet server can be minimized. Accordingly, it is possible to shorten the time of speech recognition and perform control according to immediate speech recognition.

When the voice signal section and the video signal section detected by the section detector 430 match, the character recognizing section 450 recognizes the completion type character corresponding to at least one of the voice signal and the video signal input in the matching section. The character recognition unit 450 may internally include a speech recognition engine (not shown) and a database (not shown) for speech recognition, and recognizes the completed character input within the matching period.

Specifically, based on at least one of the voice feature value and the motion feature value calculated by the character recognition unit 450 feature value calculation unit 440, the completion type character input in the matching period can be recognized.

The character recognition unit 450 can access the predetermined Internet server 410 through the network interface unit 310 and use the recognition engine 413 and the recognition data base 411 included in the Internet server 410 have. In this case, the character recognition unit 450 may not internally include a recognition engine or a database for voice recognition.

Here, the recognition database 411 has a data structure corresponding to or compared with the voice feature value or the motion feature value calculated by the feature value calculation unit 440. For example, if the voice feature value has a Mel Frequency Cepstroll Coefficient (MFCC) value, the recognition database 411 includes an MFCC value corresponding to each complete character. In addition, it may include a feature value (e.g., a motion vector value) of a video signal corresponding to each complete character.

The recognition engine 413 is an engine for determining which character the input speech signal corresponds to and recognizes the value of the speech signal or the video signal output from the speech recognition processing unit 320 (specifically, the speech characteristic value or the motion characteristic value) A value stored in the corresponding recognition database 411 is compared and detected, and it is determined whether or not the voice signal corresponds to which character.

The image display apparatus 100 may lack storage space to include a large-capacity recognition database. Accordingly, by using the recognition database 411 and the recognition engine 413 present in the Internet server 410, speech recognition can be performed more accurately. In addition, since only feature values calculated by the feature value calculating unit 440 are used in this embodiment, the amount of data signals transmitted to the Internet server 410 can be minimized. Accordingly, high-speed transmission of signals is possible, and the amount of data packets used can be minimized.

When the recognition engine 413 of the Internet server 410 is used, the character recognition unit 450 receives the recognized character from the recognition engine 413 and combines the transmitted characters to interpret the control command.

The control unit 330 performs a control operation of the image display apparatus 100 according to the control command interpreted by the character recognition unit 450.

In addition, the control unit 330 receives the recognized character transmitted in real time by the character recognition unit 450 and controls the recognized character to be output in a user interface (UI) in real time.

Specifically, the user interface unit 360 generates user interface data, and the display unit 370 displays user interface data. In addition, the control unit 330 can receive the recognized character transmitted in real time by the character recognition unit 450 and control the recognized character to be output in real time on an OSD (On Screen Display). Specifically, the OSD generation unit 355 generates the recognized character using the OSD, and the display unit 370 displays the generated OSD data. The controller 330 may control the OSD generator 355 to convert OSD data output from the user interface 360 into OSD data and output the OSD data to the display unit 370.

5 is a view illustrating a display screen output by the image display apparatus according to an embodiment of the present invention. Referring to FIG. 5, a display screen 510 or 550 is displayed when a voice signal and a video signal are inputted and the voice recognition processor 320 performs a voice recognition operation. In FIG. 5, a case where the user inputs voice signals of 'channel, null, and fill (7)' to the remote control unit 390 at predetermined time intervals is shown as an example.

Referring to FIG. 5A, a full-length character 'Chae' is input, and the voice recognition processor 320 recognizes the 'Chae' character and accordingly recognizes the recognized character 531 in the user interface data 520 ) Is displayed. A symbol 535 indicating a predetermined time interval is displayed during a period in which the input of the voice signal is temporarily stopped.

Referring to FIG. 5B, 'null', which is a one-piece type character, is input to recognize the 'null' character in the voice recognition processing unit 320, A character ('null') is displayed. Further, a symbol indicating a predetermined time interval is continuously displayed during a period in which the input of the voice signal is temporarily stopped.

6 is a view illustrating another display screen output by the image display apparatus according to an embodiment of the present invention. In FIG. 6, when the input and recognition of the voice signal to the above-described 'Chain, Null, and Chil (7)' in FIG. 5 is completed, the user interface unit 360, under the control of the control unit 330, And an OSD 610 including an OK key 621 for confirming the voice recognition and an end key 625 for terminating voice recognition without performing a control operation.

When the user operates the remote control unit 390 to input the confirmation key 621 to the control unit 330, the control unit 330 interprets the command designated by the confirmed character and executes the interpreted predetermined command. Accordingly, in the video display device 300, the channel switching operation is performed to 'channel 7'.

In FIGS. 5 to 6, when the user inputs a voice signal of 'fill', the video display device 300 is set to recognize the number '7' as an example.

7 is a view illustrating another display screen output by the image display apparatus according to the embodiment of the present invention.

Also, the control unit 330 can confirm once again whether or not to execute a predetermined command corresponding to the recognition character. Referring to FIG. 7A, the controller 330 controls the OSD to be displayed on the display screen 700, thereby confirming whether or not to execute a predetermined command corresponding to the recognized character.

In addition, the illustrated OSD itself is formed of a user interface (specifically, a GUI (Graphic User Interface) so that a user or the like can operate the remote control unit 390 and input the 'yes' key 710 to the control unit 330 Accordingly, the user can remotely control the video display device 300 to perform channel switching to 'channel 7'.

In accordance with the control of the control unit 330, the OSD 750 informing the user of the channel switching operation can be output.

Also, when the 'No' key 720 is input to the control unit 330, speech recognition may be terminated without performing a channel switching operation.

8 is a view illustrating another display screen output by the image display apparatus according to an embodiment of the present invention.

In addition, when at least one voice-like similar word exists in the recognized completed character, the control unit 330 may control the voice similar word to be outputted through the user interface. In addition, the user interface unit 360 may receive a complete character among the output at least one phonetic alphabet through the user interface. The completed character is input to the control unit 330.

Referring to FIG. 8, when a user inputs a voice signal of the character 'E', the character recognition unit 450 recognizes the characters' E ',' E 'and' have. In this case, for accurate voice recognition, the control unit 450 controls the user interface to select one of a plurality of voice similarities to be outputted.

The user or the like selects the intended letter through the user interface displayed on the OSD 810. [ Accordingly, the control unit 330 receives the selected character and performs a corresponding control operation.

9 is a view showing another display screen output by the image display apparatus according to an embodiment of the present invention.

The control unit 330 receives a voice signal (e.g., 'back' or 'back' to indicate back) for canceling the recognized character when there is an error in the recognized character, And performs voice recognition again. That is, the control unit 330 registers the voice signal (for example, 'back' or 'back') for canceling the recognized character in advance in the storage unit 380, Or " back " is inputted, the previous recognition character is canceled.

If the character intended to be recognized by the user is 'OZ', the user must input a voice signal in '' o '+ a predetermined time interval +' jet '+ a predetermined time interval'. However, if the user erroneously inputs a voice signal at "+" a predetermined time interval + '+' + a predetermined time interval ", the OSD 910 for voice recognition as shown in FIG. 9 (a) .

When the user continuously inputs a voice signal "back" + a predetermined time interval for requesting cancellation of the recognized voice signal, the control unit 330 cancels the voice signal ' And controls the OSD 920 for speech recognition as shown in FIG. 9 (b) to be outputted.

When the user again inputs the voice signal "jet + plus a predetermined time interval", the voice recognition processor 320 recognizes the alphabetical completion character 'z', and the control unit 330 recognizes The OSD 930 for voice recognition as shown in FIG.

10 is a view illustrating another display screen output by the image display apparatus according to an embodiment of the present invention.

Referring to FIG. 10, when there are a plurality of commands corresponding to word characters ('OZ' in FIG. 9) composed of recognized complete characters, the control unit 330 selects the plurality of commands It is possible to control the OSD 1010 including the user interface data to be output to the display screen 1000.

Referring to FIG. 10, when there is a command to connect to the OZ homepage, a command to connect to the LGT homepage, and the like corresponding to the voice signal 'OZ', corresponding commands are displayed on the OSD 1010 So that a user or the like can select a command among a plurality of commands displayed.

The control unit 330 selects a command through a user interface (specifically, GUI) output to the OSD 1010 to be displayed, and performs an operation corresponding to the selected command. For example, when the user selects 'command to connect to OZ homepage' and inputs the command to the control unit 330, the control unit 330 accesses the OZ homepage through the network interface unit 310, And outputs the screen to the display unit 370.

7) composed of the recognized complete character or completed character (for example, 'channel 7' as shown in FIG. 6) or recognized completion characters (for example, Quot ;, "switch to channel 7") as shown in FIG. After completing the input of the voice signal and recognizing the completion type character, the control unit 330 searches the Internet server for information related to the command corresponding to the recognized completion type character, and controls the retrieved information to be output through the user interface have. For example, as described in FIGS. 9 and 10, when the recognized completed character is 'OZ', information related to the OZ (for example, the location of the OZ service center of LG Telecom) is retrieved from the Internet server, So that the retrieved information can be controlled to be output through the user interface.

11 is a diagram illustrating a speech recognition method according to an embodiment of the present invention.

11, a speech recognition method according to an exemplary embodiment of the present invention receives a speech signal corresponding to a one-shot type character at predetermined time intervals (step 1115). While the speech signal is being input, (Step 1117).

Then, in step 1120, a speech signal section corresponding to the one-time type character is extracted (step 1125), and a video signal section corresponding to the one-time type character is extracted from the motion image (step 1120).

If the voice signal section and the video signal section coincide with each other, the completion character corresponding to at least one of the voice signal and the video signal input in the matching section is recognized (step 1130). Specifically, the signal in the matching interval is transmitted to the speech recognition engine in step 1135, and the completion character is recognized in step 1137 using the speech recognition engine.

Then, the recognized completion type character is analyzed and a command corresponding to the recognized completion type character is calculated (step 1140).

Accordingly, an instruction corresponding to the recognized complete character is executed (Step 1150).

The speech recognition method according to an embodiment of the present invention is the same as that of the image display apparatus according to the embodiment of the present invention described with reference to FIG. 3 to FIG. Therefore, detailed description is omitted.

The terminology used in the present invention is defined in consideration of the functions of the present invention, and it may vary depending on the intention or custom of a technician working in the field. Therefore, .

The present invention is not limited to the above-described embodiments and various changes and modifications may be made by those skilled in the art, which is included in the spirit and scope of the present invention as defined in the appended claims.

110: Video display device
340: remote control device
300: Video display device
310: Signal input section
311: Tuner
315: Network interface unit
320:
330:
340: a signal processor
350:
355: OSD generating unit (On Screen Display generator)
360: User Interface Unit
370: Display unit
380: Storage unit
390:
391:
393: microphone section
395:
397: Remote control interface unit

Claims (20)

A method of controlling an image display apparatus using speech recognition,
Storing a specific audio signal for deleting the recognized complete character in a storage unit;
Receiving a voice signal corresponding to a one-time type character at predetermined time intervals;
Taking a mouth-shaped motion image;
Extracting a voice signal section corresponding to the one-letter type character from the voice signal, and extracting a video signal section corresponding to the one-letter type character from the motion picture;
Recognizing the completion type character corresponding to at least one of the voice signal and the video signal input in the matched section if the voice signal section and the video signal section coincide with each other;
Deleting the recognized complete character in response to a specific audio signal stored in the storage unit; And
And performing a predetermined command including the recognized complete character. The method of claim 1, wherein recognizing the completed character comprises:
Calculating a speech feature value of the speech signal input in the matched interval; And
And recognizing the one-shot type character based on the voice feature value. 3. The method of claim 2, wherein recognizing the finished character comprises:
And calculating a motion feature value of the motion image input in the matched section. 4. The method of claim 3, wherein recognizing the one-time character based on the voice feature value comprises:
Recognizing the complete character based on the voice feature value and the motion feature value. The method of claim 1, wherein the step of receiving the voice signal comprises:
And receiving at least one of the at least one voice signal corresponding to at least one of the completed characters forming the predetermined command at the predetermined time intervals. The method of claim 1, wherein the step of extracting the video signal section
Extracting, as the voice signal section, a section from the time when the voice signal starts to be input until the start of the predetermined time interval; And
And extracting, as the video signal section, a section from a point at which the mouth shape starts moving to a point at which the mouth shape stops moving. The method according to claim 1,
Outputting information corresponding to the predetermined command including the recognized completed character or the completed character through a user interface in real time; And
Further comprising the step of performing the predetermined command. The method of claim 1, wherein recognizing the completed character comprises:
Transmitting the audio signal and the video signal input in the matched section to an Internet server; And
And recognizing the completion type character using the speech recognition engine and the speech recognition database of the Internet server. The method according to claim 1,
Outputting the voice similarity through a user interface when at least one voice similarity exists in the recognized completed character; And
Further comprising the step of the user selecting a desired completed character among the at least one voice similarity through the user interface. The method according to claim 1,
Outputting the plurality of predetermined commands through a user interface when there are a plurality of predetermined commands including the recognized completed characters; And
Further comprising receiving a predetermined command from the plurality of predetermined commands through the user interface. The method according to claim 1,
Outputting information corresponding to the predetermined command including the recognized completed character or the completed character through a user interface in real time; And
Searching the Internet server for information related to the predetermined command when the input of the voice signal is completed, and outputting the retrieved information through the user interface. A remote controller for inputting a voice signal corresponding to the one-letter type character at predetermined time intervals and photographing a mouth-shaped motion picture;
Extracting a voice signal section corresponding to the one-letter type character from the voice signal, extracting a video signal section corresponding to the one-letter type character from the motion image, and, if the voice signal section and the video signal section coincide, A voice recognition processor for recognizing the completion type character corresponding to at least one of the voice signal and the video signal inputted within a predetermined interval;
A storage unit for storing a specific audio signal for deleting a recognized character; And
And a control unit for controlling the execution of a predetermined command including the recognized complete character,
Wherein,
And deletes the recognized complete character in response to the specific audio signal stored in the storage unit. 13. The apparatus of claim 12, wherein the speech recognition processor
Calculates a voice characteristic value of the voice signal input in the matched section, and recognizes the one-finished character based on the calculated voice characteristic value. 14. The apparatus of claim 13, wherein the speech recognition processor
Calculates a motion feature value of the motion image input in the matched section, and recognizes the monotone type character based on the calculated voice feature value and the motion feature value. 13. The apparatus of claim 12, wherein the speech recognition processor
A section from the time when the voice signal is input until the beginning of the predetermined time interval is extracted as the voice signal section and a section from the time when the mouth shape starts moving to the time when the mouth shape is stopped And extracts the image signal in the video signal section. 13. The method of claim 12,
Further comprising a user interface unit for outputting user interface data including information corresponding to the predetermined command including the recognized complete character or the completed character in real time. 13. The method of claim 12,
Further comprising a network interface unit for transmitting and receiving data with at least one Internet servers,
The speech recognition processor
And transmits at least one of the audio signal and the video signal input in the matched section through the network interface unit to the Internet server and recognizes the completion type character using the speech recognition engine and the speech recognition database of the Internet server And the video display device. 17. The method of claim 16,
The speech recognition processor
And controlling the voice similarity output through the user interface unit when at least one voice similarity exists in the recognized completed character,
The user interface unit
Wherein the at least one of the at least one phonetic alphabet is selected by the user. delete 17. The apparatus of claim 16, wherein the speech recognition processor
Controlling the plurality of predetermined commands to be output through the user interface unit when a plurality of predetermined commands including the recognized completed characters exist,
The user interface unit
Further comprising a video display unit for receiving a predetermined command among the plurality of predetermined commands.

Images