JP2000165797A - Information processor and information processing method and providing medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To surely detect CM included in television broadcasting. SOLUTION: A CM candidate block detecting part divides scene change blocks as shown in (A), calculates the time of the blocks, and divides stereo blocks as shown in (B). Then, the CM candidate block detecting part judges CM candidate blocks from the scene change blocks and the stereo blocks. Next, the CM candidate block detecting part judges that the time of CM is 7 second units, or 14 second units, or 28 second units since the total time of the CM candidate blocks is 56 seconds, and the CM candidate block detecting part finally judges that the time of CM is the integral times of 14 seconds since the scene change is operated in the intervals of 14 seconds or 28 seconds.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information processing apparatus and method, and a providing medium, and more particularly to an information processing apparatus and method for detecting a commercial included in a television broadcast, and a providing medium.

[0002]

2. Description of the Related Art For example, when a recorded television broadcast is reproduced, a commercial (hereinafter, referred to as C) is required in response to a user's need to continuously watch only the main part (program).
There is a so-called VCR (Video Cassette Recorder) with a so-called CM cut function for fast-forwarding M (described as M).

[0003] The CM detection algorithm used in such a VCR is based on the features of the absolute or most CMs as follows. That is, at the start and end of each CM, 0.1 to 2.0
Features such as a silent section of about seconds, a scene change point of an image in a silent section, the required time of each CM is an integral multiple of 15 seconds, and the CM has a stereo multiplex audio multiplex mode. Is detected, that part is detected as CM.

In television broadcasting in the United States and Europe, commercials are based on the absolute condition that a black frame or a blue frame exists between the main program and the commercial.
Has been detected.

[0005]

Therefore, the conventional C
According to the M detection algorithm, for example, there is a problem that a CM that does not include the above-described features, such as a CM whose audio multiplex mode is monaural, cannot be detected.

Further, according to the conventional CM detection algorithm, when the above-described features are included in the main part, there is a problem that such a portion is detected as a CM.

Further, the characteristics of the CM used in the conventional algorithm are changed, for example, the time required for the CM becomes an integral multiple of 14 seconds, the insertion of black frames and blue frames is abolished in Europe and the United States, or When abolished, there was a problem that it became impossible to detect a CM.

In the conventional silent detection method in the CM detection algorithm, an average level of voice in a certain section is obtained, and when the average level is equal to or less than a predetermined threshold, the section is detected as a silent section. Was. For this reason, when the reception condition of the television broadcast is poor, that is, when the S / N of the audio signal is poor due to the weak electric field, the detection accuracy of the non-speech section differs when the electric field is not a weak electric field. There was a problem that CM could not be detected.

The time required for a commercial in the conventional commercial detection algorithm is measured by counting the number of frames at about 30 frames per second, but in actual broadcasting, it is about several frames. Since an error occurs, the threshold for determining the required time is provided with play corresponding to the error. Therefore, there is a problem that erroneous detection or non-detection of a CM occurs due to the play.

[0010] Further, when a commercial spot of a television station for about 5 seconds is broadcast between the commercial and the main program, the spot is similar to the commercial for the viewer, but has the above-mentioned features of the commercial. There was a problem that could not be detected because it was not.

The present invention has been made in view of such a situation, and it is an object of the present invention to reliably detect a CM included in a television broadcast.

[0012]

According to an embodiment of the present invention, there is provided an information processing apparatus, comprising: storage means for storing a detection condition corresponding to a feature of a commercial; and a commercial based on the detection condition stored in the storage means. Detecting means for detecting a change in the characteristics of the commercial, and updating means for updating the detection condition stored in the storage means in accordance with the detection result of the detecting means.

According to a fifth aspect of the present invention, there is provided an information processing method comprising: a storage step of storing a detection condition corresponding to a feature of a commercial; a detection step of detecting a commercial based on the detection condition stored in the storage step; And a updating step of updating the detection condition stored in the storage step in accordance with the detection result of the detection step.

[0014] According to a sixth aspect of the present invention, there is provided the providing medium, comprising: a storage step of storing a detection condition corresponding to a commercial feature; a detection step of detecting a commercial based on the detection condition stored in the storage step; To provide a computer-readable program for causing an information processing apparatus to execute a process including a detection step of detecting a change and an update step of updating a detection condition stored in a storage step corresponding to a detection result of the detection step. It is characterized by.

[0015] The information processing apparatus according to claim 1,
In the information processing method described in (1) and the providing medium described in (6), a detection condition corresponding to the feature of the commercial is stored, and the commercial is detected based on the stored detection condition. Further, a change in the feature of the commercial is detected, and the stored detection condition is updated in accordance with the detection result.

[0016]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described below. In order to clarify the correspondence between each means of the invention described in the claims and the following embodiments, each means is described. When the features of the present invention are described by adding the corresponding embodiment (however, an example) in parentheses after the parentheses, the result is as follows. However, of course, this description does not mean that each means is limited to those described.

According to the first aspect of the present invention, there is provided an information processing apparatus comprising: storage means for storing a detection condition corresponding to a feature of a commercial (for example, a CM candidate section detection unit in FIG. 2); Detecting means for detecting a commercial based on the condition (for example, the CM candidate section detecting unit 16 in FIG. 2), detecting means for detecting a change in the feature of the commercial (for example, the CM candidate section detecting unit 16 in FIG. 2), An update unit (for example, the CM candidate section detection unit 16 in FIG. 2) for updating the detection condition stored in the storage unit in accordance with the detection result of the detection unit is provided.

A first configuration example of a VCR to which the present invention is applied will be described with reference to FIG. Recording tuner 2
Demodulates the television broadcast RF signal input from the terminal 1 and converts the obtained video signal, audio signal, AGC signal, and signal indicating the audio multiplex mode (hereinafter referred to as audio multiplex mode signal) into Output to the CM detection circuit 3. The tuner 2 outputs a video signal and an audio signal to the delay circuit 4.

The CM detection circuit 3 determines whether or not the signal input from the tuner 2 is a CM, and outputs 0 or 1 to the switch 6 as a control signal in accordance with the result. The switch 6 is turned off when the control signal is 1, and is turned on when the control signal is 0.

The delay circuit 4 delays the video signal and the audio signal input from the tuner 2 for the time required for the processing of the CM detection circuit 3 (1 minute in the present embodiment) and outputs the delayed signal to the modulation circuit 5. I do. The modulation circuit 5 includes a delay circuit 4
The video signal and the audio signal input from are compressed and encoded by a predetermined method (for example, MPEG2 method), and further modulated by a predetermined modulation method (for example, EFM modulation). Output to the writing circuit 7 via the The writing circuit 7 records the input modulation signal on the magnetic tape 8.

The reading circuit 9 of the reproducing system reads and demodulates a signal recorded on the magnetic tape 8 in response to a command input by a user, and supplies the demodulated signal to a monitor (not shown).

The medium for recording the video signal and the audio signal is not limited to a magnetic tape, but may be an optical disk, a magneto-optical disk, a hard disk, a semiconductor memory, or the like.

FIG. 2 shows a first configuration example of the CM detection circuit 3 of FIG. In the CM detection circuit 3, the audio multiplex mode signal input from the tuner 2 is supplied to the silence threshold value determination unit 11 and the CM candidate section detection unit 16, and the video signal is supplied to the delay unit 14 and the scene change detection unit 15, The audio signal is supplied to the A / D conversion unit 12, and the AGC signal is supplied to the silence threshold value determination unit 11.

The silence threshold value determining unit 11 uses any one of the audio multiplex mode signal, the audio signal digitized by the A / D conversion unit 12, the AGC signal, and the signal input from the scene change detection unit 15 to generate silence. The threshold value used for the section detection processing is calculated and supplied to the silent section detection unit 13.

The silent interval detecting unit 13 detects a silent interval by comparing the level of the digitized audio signal input from the A / D converter 12 with the threshold value supplied from the silent threshold value determining unit 11. Then, the result is output to the scene change detection unit 15.

The scene change detecting section 15 receives two frame images (the current frame and the delay
(The previous frame delayed by one frame time (1/30 second)) to detect the presence or absence of a scene change in a silent section, and to report the result as a CM candidate section detector 16
Output to

The CM candidate section detecting section 16 binarizes the audio multiplex mode signal and the scene change information supplied from the scene change detecting section 15 for each frame, stores the binary in a built-in memory, and based on the information. In the CM candidate section, 1 is output as a control signal to the switch 6, and in other than the CM candidate section, 0 is output as a control signal. Note that the CM candidate section detection unit 1
6 stores information for the past one minute. In this case, the capacity of the built-in memory is 60 (seconds) × 30 (frames) × 2 (data) × 1 (bits). Become.

Next, the operation of the CM detection circuit 3 will be described with reference to the flowchart of FIG. This CM detection processing is started when each signal is input from the tuner 2 to the CM detection circuit 3. In step S1, the CM
The silence threshold value determination unit 11 of the detection circuit 3 outputs to the silent interval detection unit 13 a threshold value calculated in advance (described later) and used for the silent interval detection process (step S2).

Here, the silence threshold value determination processing will be described in detail with reference to the flowchart of FIG. In step S11, the A / D converter 12 converts the audio signal (analog) for a predetermined very short time input from the tuner 2 into
A digital audio signal obtained by digitizing at a predetermined sampling frequency and a predetermined quantization level (for example, FIG.
(A)) is output to the silence threshold determining unit 11. The silence threshold determining unit 11 converts the digital audio signal input from the A / D conversion unit 12 into an absolute value as shown in FIG. 5B in step S12, and outputs the absolute value of the sample in step S13. The average value of the level (FIG. 5C) is calculated.

In step S14, the silence threshold value determining section 11 compares the average value obtained in step S13 with the threshold value stored so far, and stores the smaller value as a new threshold value.

In step S15, the silence threshold determining section 11 determines whether or not the processing of steps S11 to S14 has been performed for all the channels, and determines that the processing has not been performed for all the channels. If so, the process proceeds to step S16. In step S16, the silence threshold value determining section 11 outputs a channel switching signal to the tuner 2. The channel is switched according to the channel switching signal.

Thereafter, in step S15, when it is determined that the processes of steps S11 to S14 have been performed for all the channels, the silence threshold value determination process ends. This silence threshold value determination process is repeatedly executed at predetermined intervals (for example, every 10 minutes).

It should be noted that a method of the silence threshold value determination processing may be considered other than the above-described method. For example, at the broadcast start time and broadcast end time of each channel, the audio signal is received, and a predetermined offset value Δ is added to its level n, although there is a video signal but the level of the audio signal is definitely 0. The value n + Δ may be used as the silence threshold. However, it is assumed that the broadcast start time and the broadcast end time of each channel are known.

The silence threshold may be determined using the AGC signal from the tuner 2. That is, as shown in FIG. 6, the level of the AGC signal from the tuner 2 is inversely proportional to the level of the AGC signal and the S / N of the audio signal. Therefore, if the S / N of the audio signal is estimated from the level of the AGC signal,
A threshold value corresponding to the estimated S / N can be uniquely determined.

Furthermore, since the buzz component of the audio signal depends on the luminance level of the video signal, the silence threshold may be determined using the average value of the luminance level of the video signal. For example, when the average value of the luminance level is high, the buzz component of the audio signal increases, and as a result, the level of the audio signal for a certain period of time increases. Therefore, the threshold value is set higher than usual, and the average value of the luminance level is set. Is low, the buzz component of the audio signal decreases,
As a result, the level of the audio signal decreases for a certain period,
The threshold is set lower than usual. The average value Y _A of the luminance of the video signal, the following equation is calculated at the scene change detector 15. Y _A = (ΣDij) / n × m where i = 1 to n, j = 1 to m, and Dij is
As shown in FIG. 7, coordinates (i, j) of the image corresponding to the audio signal are pixel values.

Further, the silence threshold may be determined based on the audio multiplex mode signal. That is, the level of the audio signal when the audio multiplex mode is switched from bilingual broadcasting to stereo may be set as the silence threshold.

Returning to FIG. In step S2, the silent interval detecting unit 13 determines in step S1 that the silent threshold determining unit 11
The silent section is detected on the basis of the threshold value input from. The details of the silent section detection processing will be described with reference to the flowchart in FIG.

In step S21, the A / D converter 12
Is a digital audio signal (for example, FIG. 5A) obtained by digitizing an audio signal (analog) for a predetermined minute time inputted from the tuner 2 at a predetermined sampling frequency and a predetermined quantization level. Is output to the silence threshold determination unit 11. In step S22, the silent section detection unit 13 converts the digital audio signal input from the A / D conversion unit 12 into an absolute value as shown in FIG.
In (2), an average value (FIG. 5 (C)) of the levels of the samples converted into absolute values is calculated.

In step S24, the silence threshold determining unit 11 determines whether the average obtained in step S23 is smaller than the threshold input from the silence threshold determining unit 11, and the average is smaller than the threshold. When the determination is made, the process proceeds to step S25. In step S <b> 25, the silent section detecting section 13 determines that the section is a silent section, and outputs the information to the scene change detecting section 15.

On the other hand, if it is determined in step S24 that the average value is not smaller than the threshold, the process proceeds to step S2.
Proceed to 6. In step S26, the silent section detecting unit 1
3 indicates that this section is not a silent section (it is a sound section)
And outputs the information to the scene change detection unit 15.

Returning to step S3 in FIG. In step S <b> 3, the scene change detection unit 15 detects a scene change in two frames before and after the input frame, and outputs the result to the CM candidate section detection unit 16. The details of the scene change detection process are described in FIG.
This will be described with reference to the flowchart of FIG.

In step S31, the scene change detecting section 15 determines whether or not the information input from the silent section detecting section 13 indicates a silent section, and determines whether the input information is information indicating a silent section. If it is determined that there is, the process proceeds to step S32.

In step S32, the scene change detecting section 15 calculates the correlation value E between the two preceding and succeeding frame images. Specifically, as shown in FIG. 10, an image (delayed image) input via the delay unit 14
Then, the sum of the absolute values of the pixel values of the corresponding pixels of the image (through image) input without passing through the delay unit 14 is calculated as a correlation value by the following equation. E = Σ | Dij−Sij | where Dij and Sij represent pixel values of coordinates (i, j) of the delayed image or the through image, respectively. In addition,
The correlation value increases as the degree of correlation between frames decreases, and decreases as the degree of correlation between frames increases.

Incidentally, as other methods for calculating the correlation value E, a method using a histogram of the pixel values of the delayed image and the through image, a method of dividing the delayed image and the through image into predetermined blocks, and calculating the correlation value for each block A method of calculating a difference between pixel values may be used.

In step S33, the scene change detecting section 15 determines whether or not the correlation value obtained in step S32 is larger than a predetermined threshold, and the correlation value is larger than the predetermined threshold (correlation between frames). Is smaller), the process proceeds to step S34.

In step S34, the scene change detecting section 15 determines that a scene change exists between the two frames before and after the input frame, and outputs the information to the C frame.
Output to the M candidate section detection unit 16.

Conversely, if it is determined in step S33 that the correlation value is not larger than the predetermined threshold value (the degree of correlation between frames is large), the process proceeds to step S35.

In step S35, the scene change detecting section 15 determines that there is no scene change between the two frames before and after the inputted frame, and outputs the information to the CM candidate section detecting section 16.

If it is determined in step S31 that the information is not information indicating a silent section, the information is determined to be CM information.
The signal is output to the candidate section detection unit 16 and is output to step S4 in FIG.
Return to

In step S 4, the CM candidate section detecting section 16 is based on the audio multiplex mode signal for each frame in the past minute and the information from the scene change detecting section 15, which has been binarized and stored in a built-in memory. To determine a CM candidate section.

That is, as shown in FIG. 11, the memory incorporated in the CM candidate section detection unit 16 has a stereo mode of 1, a monaural mode and a bilingual mode for an audio multiplex mode signal (Audio_Multi []). Is recorded as 0, and in the information (Scene_Change []) input from the scene change detection unit 15, a frame that is a scene change is recorded as 1 and a frame that is not a scene change is recorded as 0.

The CM candidate section detecting section 16 refers to the built-in memory and, as shown in FIG. 12A, sets a section (scene change point) for each frame (scene change point) in which a signal indicating a scene change is 1. , Scene change section 0 to 10)
Is divided, and the number of frames constituting each scene change section is divided by 30 to calculate the time of that section. In addition, as shown in FIG. 12B, the CM candidate section detection unit 16 sets a continuous frame in which the audio multiplex mode signal is 1 as a stereo section. Further, as shown in FIG. 12C, the CM candidate section detection unit 16 determines that the time of a single scene change section or the total time of a plurality of adjacent scene change sections is an integral multiple of 15 seconds, and The section that is a stereo section is determined as a CM candidate section, and 1 is output as a control signal to the switch 6 in the CM candidate section, and 0 is output as a control signal to the switch 6 in other than the CM candidate section.

The switch 6 is turned off when the control signal is 1, and is turned on when the control signal is 0. Since the switch 6 is delayed by the delay circuit 4 for one minute, the switch 6 Since the modulated video signal and audio signal that are synchronized with the control signal from the candidate section detection unit 16 are input from the modulation circuit 5,
Only the video signal and audio signal of the main part are supplied, and the video signal and audio signal of the CM candidate section are not supplied. Therefore, only the main part excluding the CM candidate section of the television broadcast is recorded on the magnetic tape 8.

Next, another operation example of the CM candidate section detection section 16 will be described. This operation example is not an absolute feature of the CM, but adds a general tendency of the CM (a scene change is performed a plurality of times in the CM) to the detection condition of the CM candidate section.

In this operation example, as shown in FIG. 13, the memory incorporated in the CM candidate section
The information (No_Sound []) indicating a silent section (output of the silent section detection unit 13) (No_Sound []) (0 for a sound portion and 1 for a silent section) is added to the information shown in 1 and recorded. In this case,
The capacity of the built-in memory is required to be 60 (sec) × 30 (frame) × 3 (data) × 1 (bit).

The CM candidate section detection section 16 refers to the built-in memory, extracts a silent section as shown in FIG. 14A, and extracts a scene change point as shown in FIG. 14B. As shown in FIG. 14 (C), a scene change section (scene change sections 0 to 10 in the case of FIG. 14 (C)) is divided for each scene change point in a silent section, and a frame constituting each scene change section By dividing the number by 30, the time of the section is calculated.

Further, the CM candidate section detection unit 16
As shown in (D), the time of a single scene change section,
Alternatively, grouping such that the total time of a plurality of adjacent scene change sections is an integral multiple of 15 seconds, and FIG.
As shown in FIG. 14, a section where the audio multiplex mode signal is 1 is extracted, and as shown in FIG. 14 (F), a section that is grouped into integral multiples of 15 seconds and is a stereo section (in this case, The sections a to c and the sections d and e) are determined as CM candidate sections. Further, the CM candidate section detection unit 16
As shown in FIG. 4 (G), C at both ends of the adjacent CM candidate section
M candidate sections (in this case, CM candidate sections a, c, d,
e), the number of scene changes is compared with a predetermined threshold (for example, 1), and the number of scene changes is larger than the threshold.
c) is determined to be a CM section. Note that CM candidate sections a and c
Is also determined as a CM section.

The CM candidate section detection section 16 outputs 1 as a control signal to the switch 6 in the CM section,
In a section other than the section, 0 is output as a control signal to the switch 6.

Next, still another operation example of the CM candidate section detecting section 16 will be described. In this operation example, when the absolute feature of the CM is changed (for example, the time of the CM is changed from an integral multiple of 15 seconds to an integral multiple of 14 seconds, etc.), the change of the feature is performed. The reference value used to determine the CM candidate section is changed.

In this operation example, the CM candidate section detection unit 16 refers to the built-in memory (FIG. 11) and
As shown in (A), a scene change section (in this case, scene change sections 0 to 10) is divided for each frame in which a signal indicating a scene change is 1, and the number of frames constituting each scene change section is 30. By dividing, the time of the section is calculated. In addition, as shown in FIG. 15B, the CM candidate section detection unit 16 sets a continuous frame in which the audio multiplex mode signal is 1 as a stereo section.

Here, the CM candidate section detecting section 16 compares the time of the stereo section with the time of a section other than the stereo section adjacent to the stereo section, and if the time of the stereo section is sufficiently short, the stereo section is determined. (For example, scene change sections 1 to 4, scene change sections 8, 9)
Is determined as a CM candidate section.

Next, the CM candidate section detecting section 16 determines that the total time of the CM candidate section (scene change sections 1 to 4) is 5
Since the time is 6 seconds, it is determined that the CM time is in units of 7 seconds, 14 seconds, or 28 seconds. Further, the CM candidate section detection unit 16 determines that the scene change is 14 seconds.
Since it is performed at intervals of seconds or 28 seconds, it is finally determined that the CM time is an integral multiple of 14 seconds.

When the above result (the CM time is determined to be an integral multiple of 14 seconds) is obtained continuously for a predetermined number of times or more, the CM candidate section detecting unit 16 sets the CM time to It is determined that it has been changed to an integral multiple of 14 seconds, and thereafter, 14
An integer multiple of seconds is used as a reference value for CM section detection.

If the absolute features of the CM are changed (such as the change of the CM time as described above and the elimination of black and blue frames in Europe and the United States),
A CM candidate section detection unit 16 outputs a reference value for CM section determination changed corresponding to the changed absolute feature from outside the apparatus.
May be supplied. In this case, the CM candidate section detection section 16 may store the reference value for CM section detection in a rewritable recording medium such as a flash memory.

Next, a second configuration example of the CM detection circuit 3 will be described with reference to FIG. This configuration example is shown in FIG.
The CM block detection unit 21 is added to the configuration example shown in FIG. The CM block detecting unit 21 uses the number of frames of a single CM candidate section detected by the CM candidate section detecting unit 16 to detect a CM block composed of a plurality of CMs sandwiched between main parts, A control signal is output to the switch 6 based on the result.

According to this configuration example, a CM block can be detected by using the following features of the CM. That is, when a plurality of CMs are broadcast continuously, each CM has a standard number of frames of 5%.
Although it has an error of about a frame, as a CM block (a plurality of continuous CMs), an error with respect to a standard number of frames is about 3 frames. For example, when four 15-second CMs are broadcast continuously, the number of frames of each CM is 450 ± 5 (= 15 × 30 ± 5).
The number of frames of the CM block is 1800 ± 20 (= 15
× 4 × 30 ± 5 × 4), but 1800 ± 3.

Next, the operation will be described with reference to FIGS. The CM block detection unit 21 determines each CM candidate section detected by the CM candidate section detection unit 16 (FIG. 17).
For (A) scene change sections 1 + 2, 3, 4, 8, and 9), the number of frames is read from the CM candidate section detection unit 16 and the number of frames of adjacent CM candidate sections is summed up. It is determined whether the error is within an allowable range (± 3). The CM block detection unit 21
If it is determined that the total number of frames is within the allowable range of error, the adjacent CM candidate section is determined to be a CM block, and if the total number of frames is not within the allowable range of error. If it is determined, it is determined that the adjacent CM candidate section is not a CM block.

In this case, as shown in FIG. 17D, the total number of frames in the adjacent CM candidate section (scene change sections 1 to 4) is 1803 (= 453 + 446 + 9).
04), and the error (3) with respect to the standard number of frames of 1800 (= 60 seconds × 30 frames) is within the allowable range. Therefore, the adjacent CM candidate sections (scene change sections 1 to 4) are CM blocks. Is determined.

On the other hand, the total number of frames in adjacent CM candidate sections (scene change sections 8 and 9) is 910 (= 4
55 + 455), and the standard number of frames 900 (=
Since the error (10) with respect to (30 seconds × 30 frames) is outside the allowable range, it is determined that the adjacent CM candidate sections (scene change sections 8 and 9) are not CM blocks.

A CM candidate section as shown in FIG.
When input from the M candidate section detection unit 16, the CM block detection unit 21 performs scene change sections 1 to 4 (referred to as CM block A) and scene change sections 7 to 10 in the same manner as in the case of FIG. 17 described above. (Referred to as CM block B) is determined to be a CM block.

Further, the CM block detecting section 21
The number of frames in the sections (scene change sections 5 and 6) that are not the CM candidate sections sandwiched between the blocks A and B are read from the CM candidate section detection unit 16, and the number of frames (150 in this case) is set to a predetermined threshold. If it is less than (for example, 300 frames (10 seconds)), it is determined that the section is not the main part but similar to the CM for the user, and is included in the CM blocks A and B before and after. That is, as illustrated in FIG. 18D, the CM block detection unit 21 determines that the scene change sections 1 to 10 are one CM block.

The CM block detecting section 21 outputs 1 as a control signal to the switch 6 in the CM block, and outputs 0 as a control signal to the switch 6 in other than the CM block.

Next, a third configuration example of the CM detection circuit 3 will be described with reference to FIG. This configuration example is shown in FIG.
Is a configuration in which a CM feature amount detection unit 31 is added to the first configuration example shown in FIG. The CM feature amount detection unit 31 has two
Frame images (video signals), digitized audio signals, information stored in a memory incorporated in the CM candidate section detection section 16 (FIG. 11), and CM candidates detected by the CM candidate section detection section 16 Section information is supplied. C
The M feature amount detection unit 31 determines whether or not the supplied information has various features of the CM, and detects a CM section based on the result. Further, the CM feature amount detection unit 3
1 outputs a control signal to the switch 6 according to the detection result.

The operation of the CM feature value detecting section 31 will be described with reference to the flowchart of FIG. This feature amount detection processing is performed for each CM detected by the CM candidate section detection unit 16.
This is executed for the candidate section.

In step S41, the CM feature value detection unit 31 initializes the evaluation value to zero. In step S42, the CM feature amount detection unit 31 counts the number of scene changes in the CM candidate section input from the CM candidate section detection unit 16, and determines that the number of scene changes is equal to a predetermined threshold (1).
(5 times per 5 seconds) or more, and if it is determined that the measured number of scene changes is equal to or more than a predetermined threshold, 1 is added to the evaluation value. Conversely, if it is determined that the measured number of scene changes is not greater than or equal to the predetermined threshold, no addition to the evaluation value is performed. This process is based on a general feature that frequent scene changes occur in commercials.

In step S42, the commercial characteristic amount detector 31 detects the periodicity of the audio signal in the commercial candidate section. This periodicity detection processing of the audio signal
It is based on the general feature that BGM is used.

In the process of detecting the periodicity of the audio signal,
This will be described with reference to the flowchart in FIG. In step S61, as shown in FIG. 22, the CM feature amount detection unit 31 detects a point where the level of the audio signal input from the A / D conversion unit 12 is equal to or higher than a predetermined threshold value and a peak (see FIG. B)). However, FIG. 22A shows an analog audio signal.

In step S62, the CM feature quantity detecting section 31 determines whether or not the peak detected in step S61 has a periodicity longer than a predetermined period (several seconds). To determine the periodicity, for example, FFT (Fast Fouri
er Transform) or simply measure the interval between peaks. If it is determined that the detected peak has a periodicity longer than the predetermined period, step S6
Proceed to 3.

In step S63, the CM feature amount detecting section 31 adds 1 to the evaluation value.

If it is determined in step S62 that the detected peak has no periodicity longer than the predetermined period, no addition is made to the evaluation value, and step S63 is skipped.

The flow returns to step S44 in FIG.
In step S44, the CM feature amount detection unit 31
The continuity of the audio signal and the video signal in the M candidate section is detected. This continuity detection processing will be described with reference to the flowchart in FIG.

In step S71, the CM feature quantity detector 31 determines whether the periodicity of the peak of the audio signal continues for more than 95% or more of the CM candidate section. For measuring the duration of peak periodicity,
The information obtained in step S43 is used. When it is determined that the periodicity of the peak of the audio signal continues for more than 95% or more of the CM candidate section, the process proceeds to step S72.

In step S72, the CM feature value detecting section 31 adds 1 to the evaluation value.

In step S73, the CM feature amount detection unit 31 calculates the difference between the pixel values of the corresponding pixels of the two preceding and succeeding images, and calculates the pixel whose difference is equal to or less than the predetermined threshold value, that is, the two pixels. Pixels that do not move between images are detected. In step S74, the CM feature amount detection unit 31
Generates a histogram of the pixel values of the pixels detected in step S73, and detects the pixel value indicating the maximum value as the representative color of the background. Note that the processes of steps S73 and S74 are repeatedly executed at predetermined intervals in the CM candidate section.

In step S75, the CM feature amount detection section 31 refers to the plurality of background representative colors detected in step S74 and determines whether the same pixel value is continuously detected as the background representative color. I do. When it is determined that the same pixel value is continuously detected as the representative color of the background, the process proceeds to step S76.

In step S76, the CM feature quantity detecting section 31 adds 1 to the evaluation value.

If it is determined in step S71 that the periodicity of the audio signal peak does not continue for more than 95% or more of the CM candidate section, no addition is made to the evaluation value. S72 is skipped.

If it is determined in step S75 that the same pixel value is not continuously detected as the representative color of the background, no addition to the evaluation value is performed, and step S76 is skipped.

Returning to step S45 in FIG.
In step S45, the CM feature amount detection unit 31 detects the reproducibility of the image. In this reproducibility detection process, for example, the CM of a ketchup for 15 seconds includes a green forest scene (3 seconds), a blue sky scene (2 seconds), a green forest scene (2 seconds), and a red ketchup. In general, the same image is reproduced in CM, which is composed of a scene (3 seconds), a green forest scene (2 seconds), and a red ketchup scene (3 seconds). It is based on features.

FIG. 2 shows the process of detecting the reproducibility of an image.
This will be described with reference to the flowchart of FIG. Step S8
In 1, the CM feature amount detection unit 31 classifies the CM candidate section by a scene change point. In step S82, the CM feature amount detection unit 31 generates a histogram of pixel values of each scene classified in step S81 in 16 gradations, and detects a pixel value indicating the maximum value as a representative color of the scene.

In step S83, the CM feature quantity detecting section 31 determines whether or not the representative color of the scene is reproduced in the CM candidate section (whether the same representative color is detected in different scenes). If it is determined that the representative color of the scene has been reproduced, the process proceeds to step S84.

In step S84, the CM feature quantity detecting section 31 adds 1 to the evaluation value.

If it is determined in step S83 that the representative color of the scene has not been reproduced, step S84 is skipped because addition to the evaluation value is not performed.

The flow returns to step S46 in FIG.
In step S46, the CM feature amount detection unit 31 detects a telop (caption) as shown in FIG. This telop detection process is based on a general feature that a telop is displayed at the end of a commercial.

The telop detection processing will be described with reference to the flowchart of FIG. In step S91, the CM feature amount detection unit 31 has the same pixel value in each frame image at a predetermined time (for example, 5 seconds) on the end side of the CM candidate section, as shown in FIG. Further, pixels adjacent to each other more than a predetermined number are detected.

In step S92, as shown in FIG. 26 (C), the CM feature quantity detecting section 31 searches for a pixel having the same pixel value as the pixel detected in step S91 and divides it into regions.

In step S93, the CM feature quantity detecting section 31 determines whether or not the area obtained in step S92 is continuous in the time direction (whether the same area exists in the preceding and succeeding frames). When it is determined that the regions are continuous in the time direction, the process proceeds to step S94.

In step S94, the CM feature amount detection section 31 detects the center of gravity of all the regions on the image, calculates the average value of the distance between the center of gravity and each point on the region, and further calculates the average value. By comparing the value with a predetermined threshold,
It is determined whether or not the areas are dense. If it is determined that the areas are dense, the process proceeds to step S95.

In step S95, the CM feature quantity detecting section 31 adds 1 to the evaluation value.

If it is determined in step S93 that the areas are not continuous in the time direction, no addition to the evaluation value is performed, and step S95 is skipped.

Returning to step S47 in FIG. In step S47, the CM feature amount detection unit 31 uses a method similar to the above-described telop detection processing to set a predetermined character (including a telop) on the image in the entire section from the start to the end of the CM candidate section. It is determined whether it exists for longer than the time. If it is determined that a character exists on the image for longer than a predetermined time, 1 is added to the evaluation value. Conversely, if it is determined that no character exists on the image for longer than the predetermined time, no addition to the evaluation value is performed. In addition,
In this process, characters are displayed in the CM,
It is based on general features.

In step S48, the CM feature amount detecting section 31 detects the ratio of the silent section other than the start point and end point of the CM candidate section. This silence rate detection processing is performed by CM
Is based on the general feature that there are few silent sections. The silence rate detection processing will be described with reference to the flowchart in FIG.

In step S101, the CM feature amount detection unit 31 reads information (FIG. 11) stored in a memory built in the CM candidate section detection unit 16, and
As shown in FIG. 28, the number of frames in a silent section other than the start point and end point of the CM candidate section is detected. Further, C
The M feature amount detection unit 31 calculates the silence rate by dividing the number of frames in the silence section detected by E by the total number of frames constituting the CM candidate section.

In step S102, the CM feature amount detecting section 31 determines whether or not the silence rate calculated in step S101 is smaller than a predetermined threshold. If it is determined that the silence rate is smaller than the predetermined threshold, step S103
Proceed to.

In step S103, the CM feature amount detecting section 31 adds 1 to the evaluation value.

If it is determined in step S102 that the silence rate is not smaller than the predetermined threshold, no addition to the evaluation value is performed, and step S103 is skipped.

The flow returns to step S49 of FIG.
In step S49, the CM feature amount detection unit 31 determines whether or not the evaluation value is larger than a predetermined threshold, and if it determines that the evaluation value is larger than the predetermined threshold, the process proceeds to step S49.
Go to 50. In step S50, the commercial characteristic amount detection unit 31 determines that the commercial candidate section is a commercial section.

If it is determined in step S49 that the evaluation value is not larger than the predetermined threshold, the process proceeds to step S51. In step S51, the commercial characteristic amount detection unit 31 determines that the commercial candidate section is not a commercial section.

In response to the result of the determination, the CM feature amount detecting section 31 outputs 1 as a control signal to the switch 6 in the CM section, and outputs the switch 6 in other than the CM section.
Is output as a control signal to the controller.

As described above, the VCR according to the present embodiment is
In FIG. 1, only the main part of the television broadcast is recorded on the magnetic tape 8, and the CM is not recorded. Therefore, when the magnetic tape 8 is reproduced, only the main part is continuously displayed.

By the way, there is a user's need to fast-forward a CM at the time of reproduction, but to record the CM. Therefore, a second configuration example of a VCR to which the present invention is applied to meet such a need will be described. This will be described with reference to FIG. In this configuration example, the switch 6 is deleted from the configuration example shown in FIG. 1, and the CM detection result of the CM detection circuit 3 is supplied to the writing circuit 7.

In the recording system of this configuration example, the writing circuit 7 records all the video signals and audio signals (the main part of the television broadcast and the CM) inputted from the modulation circuit 5 on the magnetic tape 8 and also outputs the CM detection circuit. Based on the information from 3, information on the CM (such as the temporal position of the CM) is recorded at a predetermined position on the magnetic tape 8.

In the reproducing system, when a command for reproducing only the main program (command for removing the CM) is input from the user, the reading circuit 9 is based on information on the CM recorded at a predetermined position on the magnetic tape 9. The CM is removed (fast-forward, etc.), and only the main part is demodulated and supplied to a monitor (not shown).

The application of the present invention is not limited to a VCR, but can be applied to a television receiver, a tuner, and the like.

Further, the computer program for performing each of the above processes can be provided to the user via a network medium such as the Internet or a digital satellite in addition to a medium such as a magnetic disk or a CD-ROM. it can.

[0116]

As described above, according to the information processing apparatus according to the first aspect, the information processing method according to the fifth aspect, and the providing medium according to the sixth aspect, a change in the characteristics of a commercial is detected. Since the detection condition is updated in accordance with the detection result, it is possible to reliably detect the CM included in the television broadcast.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a first configuration example of a VCR to which the present invention has been applied.

FIG. 2 is a block diagram illustrating a first configuration example of a CM detection circuit 3 of FIG. 1;

FIG. 3 is a flowchart illustrating an operation of a CM detection circuit 3.

FIG. 4 is a flowchart illustrating a process of a silence threshold value determining unit 11 of FIG. 2;

FIG. 5 is a diagram for explaining the operation of a silence threshold value determining unit 11 of FIG. 2;

FIG. 6 is a diagram for explaining the operation of a silence threshold value determining unit 11 of FIG. 2;

FIG. 7 is a diagram for explaining an operation of a silence threshold value determining unit 11 of FIG. 2;

FIG. 8 is a flowchart illustrating a silent section detection process in step S2 of FIG. 3;

FIG. 9 is a flowchart illustrating a scene change detection process in step S3 of FIG. 3;

FIG. 10 is a diagram for explaining a scene change detection process.

FIG. 11 is a diagram for explaining information recorded in a memory built in a CM candidate section detection unit 16;

FIG. 12 is a diagram for explaining processing of a CM candidate section detection unit 16;

FIG. 13 is a diagram for explaining information recorded in a memory incorporated in the CM candidate section detection unit 16.

FIG. 14 is a diagram for explaining processing of a CM candidate section detection unit 16;

FIG. 15 is a diagram for explaining processing of a CM candidate section detection unit 16;

FIG. 16 is a block diagram showing a second configuration example of the CM detection circuit 3 of FIG. 1;

FIG. 17 is a diagram for explaining processing of a CM block detection unit 21 in FIG. 16;

FIG. 18 is a diagram for explaining processing of the CM block detection unit 21 in FIG. 16;

19 is a block diagram illustrating a third configuration example of the CM detection circuit 3 in FIG.

20 is a flowchart illustrating an operation of the CM feature amount detection unit 31 in FIG.

FIG. 21 is a flowchart illustrating a process of detecting periodicity of an audio signal in step S43 of FIG. 20;

FIG. 22 is a diagram for describing a process of detecting periodicity of an audio signal.

FIG. 23 is a flowchart illustrating a continuity detection process in step S44 of FIG. 20;

FIG. 24 is a flowchart illustrating a reproducibility detection process in step S45 of FIG. 20;

FIG. 25 is a flowchart illustrating a telop detection process in step S46 of FIG. 20;

FIG. 26 is a diagram for explaining a telop transfer process.

FIG. 27 is a flowchart illustrating a silence rate detection process in step S48 of FIG. 20.

FIG. 28 is a diagram for explaining silence rate detection processing.

FIG. 29 is a block diagram showing a second configuration example of a VCR to which the present invention has been applied.

[Explanation of symbols]

2 tuner, 3 CM detection circuit, 4 delay circuit,
5 modulation circuit, 6 switch, 7 writing circuit,
8 magnetic tape, 9 reading circuit, 11 silence threshold value decision unit, 12 A / D conversion unit, 13 silence section detection unit, 14 delay unit, 15 scene change detection unit,
16 CM candidate section detecting section, 21 CM block detecting section, 31 CM feature amount detecting section

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Takamichi Mitsuhashi 6-7-35 Kita-Shinagawa, Shinagawa-ku, Tokyo F-term in Sony Corporation (reference) 5C018 AA04 AB09 AB11 AC00 HA01 HA08

Claims (6)

[Claims] 1. An information processing apparatus for detecting a commercial included in a television broadcast, comprising: storage means for storing a detection condition corresponding to a feature of the commercial; Detecting means for detecting commercials; detecting means for detecting a change in the characteristics of the commercial; and updating means for updating detection conditions stored in the storage means in accordance with the detection result of the detecting means. An information processing apparatus characterized by the above-mentioned. 2. The information processing apparatus according to claim 1, wherein the detection unit detects a change in the time of the commercial. 3. The information processing apparatus according to claim 1, wherein the detection unit detects a change in a predetermined frame image existing before and after the commercial. 4. The information processing apparatus according to claim 1, wherein the updating unit updates the detection condition stored in the storage unit using a new detection condition input from outside. 5. An information processing method for an information processing apparatus for detecting a commercial included in a television broadcast, comprising: a storage step of storing a detection condition corresponding to a feature of the commercial; and a detection step based on the detection condition stored in the storage step. A detection step of detecting a change in a characteristic of the commercial, and an update step of updating the detection condition stored in the storage step in accordance with a detection result of the detection step. An information processing method characterized by the following. 6. An information processing apparatus for detecting a commercial included in a television broadcast, comprising: a storage step of storing a detection condition corresponding to the feature of the commercial; and an information processing apparatus that stores the commercial based on the detection condition stored in the storage step. A computer that executes processing including: a detecting step of detecting; a detecting step of detecting a change in the feature of the commercial; and an updating step of updating the detection condition stored in the storing step in accordance with a detection result of the detecting step. A providing medium for providing a program readable by a user.