KR20010009415A - Recording medium storing scene change point information for navigation of A/V contents, method for automatically generating this information, and recording and/or playback apparatus for generating this information - Google Patents

Abstract

PURPOSE: A recording medium, a method for automatically generating scene turning point information and device for recording/playing back the information are provided to effectively generate the information and not to increase the charge of a hardware by monitoring a characteristic change of a video frame and automatically generating the information according to the change of characteristics. CONSTITUTION: A video sequence of a serial frame unit is inputted to a DCT(Discrete cosine transform) processing unit via a frame buffer(700). The DCT processing unit(710) performs a DCT by dividing the video data to blocks of 8X8 size. A mosaic activity processing unit(720) generates a mosaic activity adapting a statistical variance of energy values in zero frequency band. The zero frequency band is a mean brightness value of each DCT block. The entire activity processing unit(720) generates an entire activity by measuring an amount of frequency energy inside every DCT block and summing the amounts. A scene turning point generating unit(740) generates scene turning point information by discriminating a corresponding frame as a "type-1 frame" or a "type-2 frame" based on the mosaic activity and the entire activity. A navigation data generating unit(750) contains the scene turning point information in navigation data by converting to a pointer.

Description

Recording medium storing scene change point information for retrieving A / V content, a method for automatically generating this information, and a device for recording and reproducing this information {Recording medium storing scene change point information for navigation of A / V contents , method for automatically generating this information, and recording and / or playback apparatus for generating this information}

FIELD OF THE INVENTION The present invention relates to the field of digital data searching (called navigation or browsing), in particular recording media storing scene change point information for searching for audio and / or video (A / V) content, the information And a recording / reproducing apparatus for recording and reproducing this information.

In general, navigation in a device for recording and / or playing back A / V data means moving from the current A / V data position to the desired position according to the user's operation and resuming playback therefrom. It is a function to provide the reproduction according to the rearranged order when the user desires by rearranging the reproduction order of the A / V data in the order desired by the user regardless of the original recording order. Therefore, navigation data is additional information for implementing a navigation function, and is arbitrarily edited by a user or a producer, and a data string time search table for storing information corresponding to the recording position of A / V data and the reproduction time of the data. And a program table for storing the replay order relocation information.

A recording and / or reproducing apparatus is developed which stores A / V data used in digital satellite broadcasting or digital high resolution broadcasting on a one-dimensional recording medium such as a tape, and implements basic navigation functions including special reproduction such as trick playback. have. Also, to store such A / V data in a two-dimensional recording medium such as an optical disk, and to implement the function (search function) of skipping the playback order or returning to the previous screen and repeating playback according to the user's request during playback. In general, scene turning point information is used.

Therefore, the scene change point information in the apparatus for recording / reproducing navigation data for this together with the A / V data on the digital recording medium is essential information during reproduction. For example, suppose that a user has recorded all the broadcasts broadcast for one day from a specific station on a single disc. At this time, the user may feel bored to see the drama program recorded in the first part and skip to the next program. The necessary information at this time is scene change point information.

The scene change point information is pointer information indicating a specific position in the entire area of the program. The scene change point information indicates a position currently being played when the user receives a forward or backward skip command ("search command"). Points to the position where it exits and resumes playback. Scene turning point information may be recorded based on various criteria.

In the conventional digital recording / reproducing apparatus, the recording apparatus for generating scene change point information and recording it on a one-dimensional recording medium such as a tape includes a controller 102, an A / V encoder 104, and a scene change point information generator as shown in FIG. 106 and recording controller 108.

Conventionally, manual manipulation information input from a user is used to generate scene change point information. That is, the recording apparatus shown in FIG. 1 may be applied to a manual camcorder as an example, and the user is configured to repeatedly press the recording start button and the pause button to record the surrounding environments. In general, the moment when the user presses the pause button and the recording start button, it can be said that one scene change occurs. Therefore, every time such button operation occurs, the controller 102 sends a control signal to the A / V encoder 104. And the scene change point information generator 106.

The A / V encoder 104 encodes the input A / V signal according to a control signal corresponding to the recording start button provided from the controller 102, and provides the encoded A / V data to the recording controller 108, The encoding of the A / V signal is stopped in accordance with the control signal corresponding to the pause button provided from the controller 102.

Meanwhile, the scene change point information generator 106 generates scene change point information according to control signals corresponding to the recording start button and the pause button input by the user provided from the controller 102 and provides the scene change point information to the recording controller 108. . The recording controller 108 records the scene change point information generated by the scene change point information generator 106 together with the A / V data provided from the A / V encoder 104 on the recording medium 110.

As shown in FIG. 1, since a portable digital recording / reproducing apparatus such as a camcorder directly controls the recording process, the user can manually generate scene change point information by using the manual button manipulation information. In the stationary digital recording / reproducing apparatus for recording A / V data continuously transmitted, there is a problem that the generation of scene change point information cannot be effectively implemented.

That is, when recording a broadcast signal in a stationary digital recording / reproducing apparatus, a user generally applies only a simple operation of pressing the recording start button once at the start of a program and pressing the pause button when stopping recording. In this case, when the scene change point information is generated based on a user's manual operation, the scene change point information generated only points to two positions of the beginning and the end of the program. In addition, since the two moments when the user presses the recording start button and the pause button are not guaranteed to be the exact moments at which a program starts and ends, the digital recording / playback device that records the broadcast does not provide information on the scene change point information by the user's manual operation. It is not appropriate to produce.

In order to solve the above problems, it is an object of the present invention to provide a recording medium for storing the scene change point information according to the change of the video frame characteristics.

Another object of the present invention is to provide a method of automatically generating scene change point information according to a change in a video frame characteristic by tracking a change in the characteristic of a video frame.

Still another object of the present invention is to record the scene change point information according to the characteristic change of the frame by tracking the characteristic change of the video frame data, and when a search command is input, the digital recording / reproducing at high speed to the desired screen using the scene change point information. It is to provide a playback device.

In order to achieve the above objects, the recording medium according to the present invention comprises: a first area for storing audio and / or video data; And

And a second area for storing scene change point information corresponding to a point at which the video frame characteristic changes.

In the method for automatically generating scene change point information according to the present invention, there is provided a method of generating scene change point information for a digital recording and / or reproducing apparatus.

(a) analyzing a characteristic of a video frame that is continuously input and dividing it into a screen including information and a blank screen; And

(b) a scene change point generation method including generating scene change point information at the moment when the division is changed in step (a).

According to an aspect of the present invention, there is provided a device, comprising: a generator configured to inspect a characteristic of a video frame that is continuously input and generate scene change point information at a point where the characteristic changes; And

A recording control section for controlling the scene change point information and the audio / video data generated by the generation section to be recorded on a recording medium;

And a playback control unit for skipping to a desired point on the disc using the scene change point information recorded on the recording medium and reading out audio video data recorded at a desired position.

In addition, the apparatus according to the present invention is characterized in that it further comprises a playback controller for reproducing A / V data from a desired position by searching at a high speed to a desired point using scene change point information when a search command is input from a user. .

1 is a block diagram of a conventional apparatus for recording scene change point information.

2 is a block diagram of an apparatus for recording / reproducing scene change point information according to an embodiment of the present invention.

3 is a diagram illustrating a relationship between a data string time search table and an A / V data string to help understand the present invention.

4 is a diagram illustrating a relationship between cell data and A / V data strings included in a program table for better understanding of the present invention.

5 is a diagram illustrating a relationship between scene change point information and cell data included in a program for better understanding of the present invention.

FIG. 6 is a diagram for defining a type-1 frame and a type-2 frame related to a scene change point for better understanding of the present invention.

FIG. 7 is a detailed view of an automatic navigation data generator for automatically generating scene change point information of FIG. 2.

8 shows DCT block division according to the present invention.

Hereinafter, with reference to the accompanying drawings, a recording medium storing scene change point information for searching for A / V content, a method for automatically generating this information, a reproduction method using this information, and a preferred embodiment of the recording / reproducing apparatus Let's explain them.

In FIG. 2, which is a block diagram of an apparatus for recording / reproducing scene change point information according to an exemplary embodiment of the present invention, scene change point information is generated by tracking input audio and / or video data characteristic changes to include scene change point information. The controller 202, the A / V formatter 204, and the navigation data auto-generator to record to a two-dimensional recording medium such as an optical disk as a navigation data, and to search to a desired position at high speed according to a search command input from a user. 206, recording controller 208, playback controller 212, and A / V deformatter 214.

The A / V formatter 204 analyzes the content of the encoded A / V data input to the encoding unit (not shown) for encoding the original image, and divides the content into minimum recording unit blocks according to a predetermined criterion. The unit block is converted into a predetermined disc format and provided to the recording controller 208 so as to be suitable for storage in the recording medium 210.

The automatic navigation data generator 206 re-analyzes the individual minimum recording unit blocks analyzed by the A / V formatter 204 to reproduce the time required for reproducing the block and the block size information occupied in the disk space by the unit block. It is provided to the recording controller 208 as one navigation data unit, and also automatically generates scene turning point information by analyzing image frames continuously input among A / V data. In this case, the navigation data unit includes a time search table and a program table. The program table includes scene change point information generated at a point in time at which a video data characteristic of an input frame unit is converted.

The recording controller 208 records the navigation data on the disk, including the recording data (formatted A / V data) provided from the A / V formatter 204 and the scene change point information provided from the automatic navigation data generator 206. do. The controller 202 controls each block of the A / V formatter 204, the automatic navigation data generator 206, the recording controller 208, the playback controller 212, and the A / V deformatter 214.

At the time of reproduction, the reproduction controller 212 reads A / V data recorded on the recording medium 210, and records the recording medium 210 according to a control signal corresponding to a search command input by the user provided from the controller 202. A / V data recorded at a desired position is read out at a high speed by using scene change point information included in the navigation data recorded at " The A / V deformatter 214 deformats A / V data formatted in a predetermined disk format provided from the reproduction controller 212 and outputs deformatted A / V data.

For convenience of description, the recording controller 208 and the reproduction controller 212 are separately configured, but may be constituted by one block, which is generally referred to as a codec. In addition, the controller 202 may be referred to as a detector, and the navigation data generator 206 may be referred to as a generator.

FIG. 3 is a diagram showing a relationship between a data string time search table and an A / V data string recorded on a disk for better understanding of the present invention. The recording of A / V data may occur only once for all the data areas of a single disk, but in general, several recording processes are performed. Therefore, it is necessary to distinguish each recording process according to the recording order, so that different data string numbers are assigned and arranged so that each of them can be managed as one independent data string.

In addition, each data string is sequentially divided into minimum recording unit blocks according to a reference according to the reproduction time or a reference on the recording space for the efficiency of the navigation function and sequentially recorded in one data string. In the data string time search table shown in FIG. 3, for example, the A / V data string number is 2, the third minimum recording unit block of the second data string has a reproduction time of 3 seconds, and the block is located on disk space. It can be seen that the area occupied becomes one sector.

Next, a program table distinguished from the above data string time search table is shown in FIG. The program table is intended to provide a function of continuously reproducing the respective parts of different A / V data strings recorded at arbitrary positions on the disc into one logical program independently of the recording order of the A / V data strings.

One cell serves as a pointer to part or all of one A / V data string. For example, cell 6 is an A / V data string number 2 and a pointer indicating a section between 2 and 17 seconds, and a process of searching for a corresponding position on a disk using this cell information is shown in FIG. 4. . As a result, one program table composed of a plurality of cells serves as a playlist in which portions of content belonging to a plurality of data strings are grouped in an arbitrary order.

FIG. 5 is a diagram for describing a program reproduction process according to each scene change point information when one cell includes a plurality of scene change point information in the program table structure shown in FIG. 4. For example, the first scene change point information included in the cell M indicates the end of the first section of the cell M in the program M, the second scene change point information indicates the end of the second section of the cell M, and the last The kth scene change point information indicates the first portion of the fourth section of cell M.

Assuming that the second section is being played in the cell M of the current program N, when a command for skipping to the next scene is input from the user, the scene transition point information closest to the omnidirectional direction from the current playback position is released from the current playback position in the program. Find it. That is, the program playback will skip to the beginning of the third section, which is the location of the second scene change point information. On the contrary, if the user inputs a command to skip to the previous scene at the same current playback position as above, the playback state will skip the current playback position and switch to the first position of the second section indicated by the first scene change point information.

In performing a transition, the transition point is a point where the user feels that a phase transition occurs in the program, rather than an arbitrary point in the program, for example, where an advertisement is inserted in the middle of a movie. It is appropriate to point to the point at which a new conversation begins to develop between characters. The method and apparatus for automatically generating scene change point information proposed by the present invention is to maximize the convenience of using a digital recording / reproducing apparatus on the user side by automating the process of finding such a location.

FIG. 6 is a diagram for defining a type-1 frame and a type-2 frame related to a scene change point for better understanding of the present invention.

The appearance of a blank image in the broadcast reminds viewers of a program change or phase change within a program. Therefore, in order to provide the user with the convenience of navigation by capturing such a transition point when a video device that records an image received from a broadcasting station, each frame is analyzed by analyzing the sequence of video frames at the input stage. 2 frames) and general screen (ad-1, movie, inter-advertisement, etc.) frame (hereafter type-1 frame).

The type-1 frame includes clear information in the area, and when the graph 640 shows three-dimensionally, the shape of the graph resembles a rugged mountainous terrain due to its curved shape. On the other hand, in the type 2 frame, no information is included in the area, and the graph 650 showing the three-dimensional image resembles a flat plain terrain containing some noise. Here, the severity of the curvature represented by the three-dimensional graph of the image data is defined by the term "activity" of the image. In addition, an image having high "activity" is defined as clearly including meaningful information in the area, and an image having low "activity" is defined as an empty image having no distinct information in the area.

FIG. 7 is a detailed view of the automatic navigation data generator 206 which automatically generates the scene change point information of FIG. 2. The apparatus of FIG. 7 is a DCT operation processor 710 for block DCT by inputting an image frame sequence among A / V data, and a mosaic activity operation for processing mosaic activity and global activity with a DCT block generated by the DCT operation processor 710. Scene change point generation unit for generating a scene change point by using the mosaic activity and the global activity generated in the processing unit 720 and the global activity calculation processing unit 730, the mosaic activity calculation processing unit 720 and the global activity calculation processing unit 730 ( 740 and a navigation data generator 750 for generating navigation data including scene change point information generated by the scene change point generator 740.

First, an image sequence of a unit of frame received through an input terminal is input to the DCT operation processor 710 via a frame buffer 700.

The DCT computation processor 710 divides the image data [B] _{n, m} input for each frame into blocks of 8 × 8 size to perform DCT. The mosaic activity calculating unit 720 generates a mosaic activity taking a statistical variance of energy values of a zero frequency band, which means an average brightness value for each DCT block processed by the DCT calculating unit 710.

The global activity calculation processing unit 730 measures the amount of high frequency energy present in each DCT block processed by the DCT operation processing unit 710 for all DCT blocks, and generates a global activity (LAM) that is the sum of these values. .

The scene change point generating unit 740 is based on the mosaic activity and the global activity (LAM) generated by the mosaic activity calculating unit 720 and the global activity calculating unit 730, and whether the corresponding frame is a "type-1 frame" or "type-". After determining whether it is 2 frames, scene change point information is generated. The navigation data generation unit 750 converts the scene change point information generated by the scene change point generation unit 740 into a pointer indicating an appropriate position of the encoded A / V data and includes it in the navigation data.

Next, a preferred embodiment of calculating the mosaic activity and the global activity of FIG. 7 will be described with reference to the drawing showing the DCT block division shown in FIG. 8.

First, in order to calculate the block DCT, one frame of the original image of the input terminal is divided into blocks of 8 × 8 size. That is, an image having a size of 8M and a width of 8N is divided into N × M blocks. For example, suppose that int (x): the largest integer not greater than x (e.g. int (5.1) = 5) for real x, the original leftmost pixel is the original image (1,1) and the rightmost bottom pixel is the original image ( 8N, 8M), the pixel represented by the original image (34,84) is int ((34 + 7) / 8) from the top, and the int ((84 + 7) / 8) th block from the left is [b] _{5 , Is} the (2,4) th pixel of ₁₁ , and this pixel value is represented by [b] _5,11 (2,4) Next, after block division is completed, DCT is performed according to the equation defined in Equation (1).

On the other hand, for each 8x8 block on which DCT has been performed, the "block activity" for measuring the high frequency energy in the block is calculated as in Equation (2).

Next, "global activity", which is a high-frequency energy measure for the entire image, is calculated according to Equation 3 by adding up each "block activity" obtained from Equation 2.

In addition, as shown in Equation 4, the "mosaic activity" statistically processes the distribution of the DC energy, that is, the energy value in the zero frequency band, which is the average brightness value for each block, and takes the variance of the distribution.

In other words, the "global activity" value represents the absolute sum of the high-frequency energy of each block, and the "mosaic activity" represents the degree of change in the brightness value between blocks, so the scene change point generator 740 of FIG. The image frame is classified as a "type-2 frame" if it does not exceed the threshold set for each, and if the one of the two "activity" values exceeds the threshold for each of those values, the image frame is classified as a "type-1 frame." Classify as

The present invention can be applied not only to an apparatus for recording / reproducing a high density recording medium such as an optical disk, but also to an apparatus for recording / reproducing an input broadcast signal on a one-dimensional recording medium, and the method and apparatus of the present invention are variously modified and equivalent. Examples are possible.

As described above, the present invention does not decode the A / V data, but merely monitors the characteristic change information of the video data and automatically generates scene change point information. Therefore, the user needs to perform additional operations. In addition, since the present invention can be sufficiently implemented using the DCT operator, which is an arithmetic operation unit that is basically implemented in the MPEG data compression and / or decompression unit, an effective scene change point without burdening hardware for the entire system. There is an effect that can implement the generation of information.

Claims (10)

A first area for storing audio and / or video data; And And a second area for storing scene change point information corresponding to a point where a frame characteristic of the video data changes. The audio / video data of claim 1, wherein the audio / video data is recorded in the form of a data string consisting of minimum recording unit blocks, and the scene change point information is stored as additional information, and the time required for reproducing the corresponding block is included in the additional information. And a data table time search table including block size information occupied by the corresponding block on the disk space, and a program table for continuous playback in a logical program independent of the recording order of the A / V data strings. Is stored in a program table. 2. The recording medium of claim 1, wherein the video frame characteristic comprises a degree of conversion between predetermined unit blocks in a frame or an absolute sum of high-frequency energy of each block. A method of generating scene change point information for a digital recording and / or playback device, the method comprising: (a) analyzing a characteristic of a video frame that is continuously input and dividing it into a screen including information and a blank screen; And and (b) generating scene change point information at the moment when the division is changed in step (a). The method of claim 4, wherein the step (a) Dividing the video frame data into blocks of a predetermined size; Detecting energy for each band through frequency analysis for each block divided in the above process; The magnitude of the high frequency energy is measured among the detected energy distributions for each band, and the image having a larger value than the threshold value is set as the information image, and the image having a smaller value than the threshold value having the measured value may have a blank image. A method for creating a scene change point, which includes a process of setting it to one. The method of claim 5, wherein the frequency analysis of each block comprises measuring energy of bands of blocks appearing in both horizontal and vertical directions through a discrete cosine transform (DCT). . Step (a) is Dividing the video frame data into blocks of a predetermined size; Analyzing the spatial distribution of the average brightness value of each divided block in the above process and statistically analyzing the degree of change represented by the magnitude of the value; And setting a larger image as the information image than the threshold value of the change analyzed in the process and setting the image having a smaller value as the possibility of a blank image. 7. The method of claim 6, wherein the average brightness value of each block is a coefficient when the frequency is 0 in both the horizontal and vertical directions among coefficients obtained through the discrete cosine transform for the block. In a digital recording / playback device: A generation unit for inspecting a characteristic of a video frame that is continuously input and generating scene change point information at a point where the characteristic changes; And A recording control section for controlling the scene change point information and the audio / video data generated by the generation section to be recorded on a recording medium; And a playback control section for skipping to a desired point on the disc using the scene change point information recorded on the recording medium to read audio video data recorded at a desired position. The method of claim 9, wherein the generation unit A DCT processor for DCT dividing the video data input for each frame into blocks having a predetermined size; A first activity processor for generating a first activity by calculating a statistical dispersion of energy values of zero frequency bands of each DCT block processed by the DCT processor; A second activity processor which measures the amount of high frequency energy present in each DCT block processed by the DCT processor for all DCT blocks and adds these values to generate a second activity; A scene change point generator for generating scene change point information by determining characteristics of the frame based on the first activity and the second activity generated by the first activity processor and the second activity processor; And a navigation data generator converting the scene change point information generated by the scene change point generator into a pointer indicating an appropriate position of the encoded audio video data and including the same in the navigation data.