KR100474916B1 - Apparatus and method for generating thumbnail image - Google Patents

Abstract

The present invention relates to a thumbnail image generating apparatus and method among functions provided by a personal video recorder (PVR) or other digital video database system installed in a digital TV.In particular, in the case of HD video, one thumbnail pixel per sub block. In the case of SD video, four thumbnail pixels per sub-block are generated through interpolation to solve the problem of structural thumbnail image size mismatch caused by the format of the stream, and to generate thumbnail images of consistent size. In addition, a relatively large thumbnail image can be generated while preventing excessive resolution loss during thumbnail generation. In addition, in order to reduce the size of memory required when performing resizing operation for SD video, the DC value of the I picture is divided by a predetermined unit without performing a resizing operation on the entire thumbnail image, and is independent of each unit. By resizing and applying the same scaling factor in the horizontal and vertical directions in this process, it is possible to maintain the aspect ratio of the original image.

Description

Apparatus and method for generating thumbnail image}

The present invention relates to an apparatus and method for generating a thumbnail image among functions provided by a personal video recorder (PVR) or other digital video database system mounted on a digital TV.

In general, a thumbnail image represents a specific frame of a video smaller than its original size for the purpose of browsing or searching. These thumbnail images are mainly composed of content summary, key frame generation, and story board of stored video among additional functions provided by PVR or digital video database system installed in digital TV. , And program guides for stored video media. In particular, the thumbnail image is used in various fields, such as using a thumbnail image as a representative image required to effectively provide content information of a video in an additional function such as a storyboard or a program guide in the PVR.

A conventional method of generating a thumbnail image is a method using a DC value of an I picture among frames of video encoded by MPEG. This method does not require a variable length decoder (VLD) for playing the entire video frame, and does not require scaling operations for further reduction to generate thumbnail images, greatly reducing the hardware burden and reducing the horizontal. And an image reduced by 1/8 in the vertical direction, respectively. That is, a relatively simple method has the advantage of generating a thumbnail image corresponding to the size of 1/64 of the original image without burden on the hardware.

In general, however, MPEG video streams that can be viewed on a digital TV have various formats such as HD (High Definition) and SD (Standard Definition), and the resolution of each frame of the video also varies according to the various formats.

However, the conventional method for generating a thumbnail image using a DC value generates a thumbnail image corresponding to 1/64 of the original image regardless of the format of the input video. For example, thumbnail images of HD video having a resolution of 1920x1080 and SD video having a resolution of 720x480 have a resolution of 240x136 and 90x60, respectively.

Therefore, the size of the thumbnail image may vary by more than 6 times depending on the format of the video. This difference causes a problem in that the arrangement of the thumbnail image is defined differently according to the video format in an actual application field. In other words, the display map for displaying multiple thumbnail images on the screen should be defined differently for each format, and the device that stores the thumbnails, for example, the memory map that defines the storage space in memory, will also be used for each video due to this size mismatch. There is a problem that needs to be defined differently depending on the format. In addition, in the case of SD class, thumbnail images are too small to clearly convey information to the user when displaying them.

If a size of thumbnail images having different sizes is to be matched by using a resizing operation, an amount of memory corresponding to a thumbnail image of the original size and a resized thumbnail image is required. In addition, since mathematical operations and graphic functions are required for sophisticated thumbnail resizing, corresponding hardware is required. In addition, since the aspect ratio of the original image is different according to the video format, a problem may occur due to distortion of the aspect ratio when the size is exactly matched using resizing. For example, in the case of a thumbnail image having a size of 90x60 extracted from an SD video, if it is exactly matched to the size of 240x136, there may be a visual distortion problem due to the difference in the size of the horizontal and vertical scaling.

In order to solve this problem, in the conventional method, one thumbnail image pixel is extracted per block in case of SD-level video, and one thumbnail image pixel is extracted in macroblock instead of block in case of HD-quality video. The size difference of thumbnail images of SD video is solved. This method has the advantage of matching the size of thumbnails generated in video with different formats with relatively simple hardware without post-processing operations such as resizing. However, in this method, the size of the thumbnail image is based on the SD class, and in the HD class, thumbnail images corresponding to the size of 1/256 of the original image (the size of 1/16 each in the horizontal and vertical directions) are generated. There are disadvantages in that the resolution loss is relatively large and complex images are not easily delivered and thumbnails are not visually smooth.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to provide a thumbnail image generating apparatus and method for the user to recognize the contents in a consistent size regardless of the format of the video stream. Is in.

Another object of the present invention is to provide a thumbnail image generating apparatus and method for compensating for inconsistency in size of a thumbnail image of a SD class thumbnail based on a size of a thumbnail image based on a HD class.

An apparatus and method for generating a thumbnail image according to the present invention for achieving the above object generates a thumbnail image using a DC value of an I picture, but generates a thumbnail image enlarged by an additional 4 times for an SD video stream. Characterized in that.

An apparatus and method for generating a thumbnail image according to the present invention for achieving the above object is characterized in that a process of performing a resizing operation on an I picture of SD video is performed independently for each slice.

To this end, a thumbnail image generating apparatus according to the present invention includes a DC decoder for parsing only a DC value of an I picture from an input video stream, a slice buffer for storing the DC value output from the DC decoder in units of slices, If the image including the DC value output from the DC decoder is HD, the DC value is stored in the slice buffer as it is. If the SD level, the DC value of sub-blocks included in each macro block of the image is interpolated by N. ² (N is a natural number of 2 or more) enlarged to control to store in the slice buffer, and comprises a buffer controller for outputting a DC value stored in the slice buffer to a thumbnail dedicated memory to generate a thumbnail image do.

If the input image is SD, the DC values of the four luminance blocks included in each macro block of the image are enlarged twice in the horizontal and vertical directions through linear interpolation, and the Cb and Cr blocks are respectively expanded. The DC values for the control are enlarged in the horizontal direction by linear interpolation and then controlled to be stored in the slice buffer in the slice unit.

The buffer controller transmits DC values for the Cb and Cr blocks stored in the slice buffer to a thumbnail dedicated memory, and controls the read points of the slice buffer to the Cb and Cr blocks corresponding to one thumbnail image line from the slice buffer. It is characterized in that for transmitting the overlapping DC values.

The buffer controller transmits the value of the luminance region in which the interpolated values are stored using only the DC value of the current slice to the thumbnail dedicated memory at the end of the current slice, and the luminance is stored in the interpolated values using the DC values of the current slice and the next slice. The value of the region is transmitted to the thumbnail dedicated memory at the end of the next slice.

The buffer controller does not store the value of the last luminance region in the slice buffer in the case of the last slice, and duplicates the value of the immediately previous luminance region by a read point control of the slice buffer when transferring to a thumbnail dedicated memory. It features.

The thumbnail image generating method according to the present invention includes parsing only a DC value of an I picture from an input video stream, and if the image including the DC value input in the parsing step is HD, it is included in each macro block of the image. Storing DC values of all subblocks in slice units in the slice buffer; and if the image including the DC value input in the parsing step is SD, the subblocks included in each macroblock of the image N (N is a natural number of 2 or more) times enlarged in the horizontal and vertical directions through interpolation, and storing the DC value in the slice buffer in slice units, and when the slice including the macroblock ends, stored in the slice buffer. And transmitting the DC values to a thumbnail dedicated memory.

Other objects, features and advantages of the present invention will become apparent from the following detailed description of embodiments taken in conjunction with the accompanying drawings.

Hereinafter, with reference to the accompanying drawings illustrating the configuration and operation of the embodiment of the present invention, the configuration and operation of the present invention shown in the drawings and described by it will be described as at least one embodiment, By the technical spirit of the present invention described above and its core configuration and operation is not limited.

The present invention is characterized in generating a thumbnail image using the DC value of an I picture, but generating a thumbnail image enlarged by an additional 4 times for an SD video stream. For this purpose, one thumbnail pixel is extracted per block in case of HD video, and four thumbnail pixels are generated per block in case of SD video. That is, in the case of SD video, a simple linear interpolation method is applied to each DC value, and as a result, four thumbnail pixels are generated per block by performing twice the resizing operation in the horizontal and vertical directions, respectively.

In addition, the thumbnail image generating apparatus according to the present invention performs a process of performing a resizing operation on an I picture of SD video, independently for each slice. That is, internally, there is a buffer to store data corresponding to slices, and DC values are received in units of slices, and double resizing operations are performed in the horizontal and vertical directions, respectively. As a result, four lines of thumbnail data per slice are transferred to the thumbnail dedicated memory.

1 is a block diagram showing an embodiment of a thumbnail image generating apparatus according to the present invention, a DC decoder 101 for parsing only a DC value of an I picture from a video input, and a slice buffer for storing the parsed DC value. And a DC value stored in the slice buffer 103, and a DC value stored in the slice buffer 103 to a thumbnail dedicated memory through a thumbnail memory interface 104, and controls the overall flow. It consists of a buffer controller 102.

That is, the buffer controller 102 can store data stored in the buffer 102 through a thumbnail memory interface 104 in a specific location of the thumbnail dedicated memory. Here, the thumbnail dedicated memory may be provided separately or may share a system memory. If the system memory is shared, the thumbnail memory interface can also use the system memory interface.

The DC decoder 101 has a structure that can parse only the DC value of the I picture in a very simple structure compared to the VLD of a general MPEG video decoder. That is, since the purpose is to generate a thumbnail image, there is no need to decode the P picture and the B picture, and the structure of the variable decoding table and hardware required to obtain other AC values is removed.

That is, the present invention generates thumbnail data in the slice unit of an I picture in the DC decoder 101 under the control of the buffer controller 102, stores it in the slice buffer 103, and then inputs all data corresponding to the current slice. If so, it will be transferred to the thumbnail dedicated memory. At this time, when defining the macroblock address MBA and the maximum number of macroblocks included in one slice as MB_WIDTH, the horizontal position MBX of each macroblock is MBA% MB_WIDTH. Therefore, when the MBX becomes MB_WIDTH-1, the last macroblock of the slice is input. When all DC values of the macroblocks are input, the data of the slice buffer 103 is transferred to the thumbnail dedicated memory.

The size of the slice buffer is 1800 bytes, and the basis thereof will be described below with reference to other drawings.

At this time, the generation method of the thumbnail image is configured differently depending on whether the current video stream is HD or SD.

FIG. 2 is a flowchart illustrating an example of a method of generating a thumbnail image when an input video stream is HD-class, and extracts one thumbnail pixel per block.

Here, assuming that the color format is 4: 2: 0, the macro block is composed of six sub blocks, that is, four Y component blocks and one Cb and Cr component blocks. Cb and Cr are chrominance components and represent the color of the pixel when the pixel is formed by the Y component, and thus four thumbnail pixels are generated per actual macroblock.

That is, when a macro block of a specific slice starts (step 201), all the sub blocks included in the macro block (for example, when the color format is 4: 2: 0, four Y component blocks, one Cb, and Cr) The DC value for the component block) is parsed and stored in slice buffer 103 (step 202).

Then, it is determined whether the slice is finished (step 203). As described above, the end of the slice can be known by checking whether the horizontal position MBX of each macroblock is at the time of MB_WIDTH-1. If the MBX becomes MB_WIDTH-1, the last macroblock of the slice is input. Therefore, the buffer controller 102 reads the DC value of the macroblock corresponding to the slice from the slice buffer 103 and provides a thumbnail memory interface. Via 104, transfer to the thumbnail dedicated memory (step 204).

On the other hand, if it is determined before the end of the slice in step 203, the process returns to step 202 after receiving the next macro block (step 205).

That is, since the resizing is not performed on the thumbnail image of the HD video, the number of actual DC values input to the slice buffer 103 and the number of DC values stored in the slice buffer 103, that is, a thumbnail dedicated memory. The number of DC values transmitted matches. Therefore, assuming that the color format is 4: 2: 0, since there are four Y component blocks and one Cb and Cr component block per macroblock, each slice has 6 * MB_WIDTH DC coefficients.

For example, in the case of HD video having a size of 1920x1080, each slice includes 120x6 = 720 DC coefficients. As a result, a 720-byte slice buffer is required to store one slice and two buffers must be stored during transfer (ie buffer reads), and the DC coefficients of the next slice must be stored in the buffer (ie buffer writes). (double buffering) is required, and as a result, a slice buffer 103 having a size of 1440 (= 720 * 2) bytes is required for processing HD video.

3 is a view illustrating an embodiment of a configuration of a slice buffer 103 for HD-quality video of a thumbnail image generating apparatus according to the present invention, wherein a map of the buffer separates a Y component, a Cb component, and a Cr component. The data corresponding to each component can be stored in a separate area in the thumbnail dedicated memory.

In the buffer map of FIG. 3, 240 bytes and 120 bytes, which are the maximum size of each component region, are based on the size of 1920x1080. As a result, two lines of the Y component, one line of the Cb and Cr components of the thumbnail image are generated in one slice. Finally, the size of the thumbnail image of the HD video having a size of 1920x1080 is 240x136.

FIG. 4 is a flowchart illustrating an example of a method of generating a thumbnail image when the input video stream is SD-class. The number of DC values to be stored in the slice buffer 103 is actually input to the DC decoder 101. 4 times the number of DC values.

That is, in the case of SD video, it is necessary to generate twice as much data in the horizontal and vertical directions with respect to the input data and store it in the slice buffer 103. For example, in the case of an SD video having a size of 720x480, if a simple calculation is applied, the Y component 180x4 bytes and the Cb and Cr components 180 (in order to store the result of resizing the data corresponding to one slice) A slice buffer 103 of a total size of 1080 bytes is needed, including = 45 * 4) bytes.

At this time, in the process of resizing the size of the thumbnail image of the SD-class video in the horizontal and vertical directions twice, the Y component is enlarged twice by using linear interpolation in the horizontal and vertical directions. However, the Cb / Cr component performs a 2x magnification operation by linear interpolation in the horizontal direction and a 2x magnification operation by duplicate rather than interpolation in the vertical direction. That is, in the case of the Cb / Cr component, since only one line is repeated, only the portion in which the Cb / Cr component is stored by the read pointer control of the slice buffer 103 when transferred to the thumbnail-only memory without actually storing the same in the buffer 103. You can read it again. Therefore, in the case of the Cb / Cr component, only a 90-byte buffer is required. Therefore, the slice buffer 103 needs to have a size of 900 bytes rather than 1080 bytes to store one slice.

That is, referring to FIG. 4, when a macro block of a specific slice is started (step 401), the Y component blocks of the macro block, for example, the DC values of sub blocks 0 to 3, respectively, are doubled in the horizontal and vertical directions through linear interpolation. Resizing is performed and stored in the slice buffer 103 (step 402), and double resizing in the horizontal direction is performed only through linear interpolation for the DC values of the Cb and Cr blocks of the macro block, for example, the sub blocks 4 and 5. To store in the slice buffer 103 (step 403).

Then, it is determined whether the slice is finished (step 404). As described above, the end of the slice can be known by checking whether the horizontal position MBX of each macroblock is at the time of MB_WIDTH-1. If it is determined in step 404 that the slice is finished, the buffer controller 102 reads the DC value of the macro block corresponding to the slice from the slice buffer 103 and transmits the DC value of the macro block to the thumbnail dedicated memory through the thumbnail memory interface 104. 405). On the other hand, if it is determined in step 404 before the end of the slice, the process returns to step 402 after receiving the next macro block (step 406).

5 illustrates a general conceptual diagram of a 2x magnification process of an image by linear interpolation. In FIG. 5, the portion indicated by the grid hatching shape uses the pixel value of the original image as it is, and the portion indicated by the diagonal hatching shape corresponds to the arrow (that is, 1/2 or 1/4). The interpolated value obtained by multiplying the weight value by the pixel value of the original image is used. In this case, the pixel value of the boundary part uses the pixel value of the original image that can be used with an exception, and in some cases, the average of the two pixels or the original image value is duplicated as it is.

FIG. 6 illustrates a more specific embodiment, and shows an embodiment of a resizing process for thumbnail data of SD video of a thumbnail image generating apparatus according to the present invention. The portion shown in FIG. 6 represents a part of the entire data. The DC value of the current K-th slice is input to expand the DC value in the horizontal and vertical directions through linear interpolation, and the thumbnail image data corresponding thereto is slice buffer 103. After storing in the K-th slice, the thumbnail image data stored in the slice buffer 103 is transferred to a thumbnail dedicated memory.

In this case, all data corresponding to five lines shown in FIG. 6 are stored in the slice buffer 103, and the data inside the box indicated by the dotted lines is transferred to the thumbnail dedicated memory at the end of the current slice (eg, the K-th slice). However, the data inside the box marked with a solid line is finally determined using the data of the current slice (eg, Kth slice) and the next slice (eg, K + 1th slice). When sent.

That is, in the case of an SD image having a size of 720x480, 45 macroblocks exist in one slice, and thus a buffer having a size of 180x4 + 180 = 900 bytes is required for each slice. This requires (45x4) x4 = 720 bytes because of the double resizing in the horizontal and vertical directions for the Y component, and (45x2) x2 = 180 bytes due to the double resizing in the horizontal direction for the Cb and Cr components. Because it is necessary.

Of these, the amount transferred to the thumbnail-only memory at the end of the current slice is 720 bytes (ie Y 540 bytes + Cb 90 bytes + Cr 90 bytes) generated using only the current slice (ie, ▩, ), Combined with the Y component of the previous slice (e.g., K-1th slice) and the current slice (e.g., Kth slice), is a total of 900 bytes. Then, 180 bytes of incomplete data consisting of subblocks 2 and 3 of each macroblock of the current slice must be transmitted after the next slice (eg, K + 1th slice) is applied to complete thumbnail data (ie, When the next slice (eg, K + 1th slice) ends, it is transferred to the thumbnail dedicated memory.

In this case, considering double buffering to generate a thumbnail image of the SD video, a slice buffer 103 having a size of 900x2 = 1800 bytes is required. This value is larger than 1440 bytes required for generating thumbnail images of HD video. Therefore, in the thumbnail image generating apparatus according to the present invention, a slice buffer 103 of 1800 bytes is required. In this case, since the first slice of the image does not have a previous slice, the transmitted data is 720 bytes. In the case of the last slice, since there is no next slice, the data corresponding to the last line of the thumbnail is generated by overlapping the data corresponding to the previous line. Therefore, the transmitted data is 1080 bytes.

FIG. 7 illustrates an embodiment of a configuration of a slice buffer for SD video of a thumbnail image generating apparatus according to an embodiment of the present invention, in which Y regions 1 to Y regions 4 of FIGS. 6 and 7 have the same meaning.

In FIG. 7, when DC values of respective macroblocks are input, the DC values of sub blocks 0 to 3 (that is, blocks of Y components) are further calculated by linearly interpolating values in the horizontal and vertical directions to slice the slices. In the buffer 103, the linearly interpolated values in the horizontal direction are further calculated and stored in the slice buffer 103 with respect to the DC values of sub blocks 4 and 5 (ie, blocks of Cb and Cr components). . That is, in the SD video of 720x480, DC data inputted to the thumbnail image generating device per slice is Y 45x4 + Cb 45 + Cr 45 = 270 bytes, and additionally Y 540 bytes + Cb 45 bytes + Cr 45 bytes = 630 bytes An amount of data is generated by linear interpolation (for example, average interpolation) and stored in the slice buffer 103.

In FIG. 7, a portion indicated by an oblique hatching shape is an original DC value, and an portion denoted by INT is an interpolation value calculated by averaging. That is, when the DC value is transmitted from the DC decoder 101, the buffer controller 102 determines the value of the portion denoted by INT by performing interpolation by an average operation using the original DC value. As a result, when the DC value is input, the original thumbnail image data and the duplicate thumbnail image data are respectively recorded in the portion indicated by the diagonal hatching and the portion indicated by the DUP of the slice buffer 103, and the next clock is INT. The interpolated thumbnail image data is recorded in the portion indicated by. In this case, when the two DC values are used, the addition operation is performed by reading only the upper 7 bits of each DC value. When the four DC values are used, the addition operation is performed by reading only the upper 6 bits of each DC value. You can also use the method.

In addition, in the case of using a block using the DC value of the next slice, that is, the DC value starting with N of the part indicated by Y region 4 or the part indicated by INT, the DC value of sub-block 0 or 1 of the macro block of the next slice is Because it must be entered, the corresponding data is transferred to the thumbnail-only memory at the end of the next slice.

When the slice is finished by the above-described process, the data stored in the slice buffer 103 are sequentially transferred to the thumbnail dedicated memory. The order in which data is stored in the thumbnail-only memory must match the image that is finally displayed. In this case, Y area 1 to 3 are transmitted in order at the end of the current slice, but data corresponding to Y area 4 should be transmitted first at the end of the next slice. Cb area and Cr area are overlapping data, so it is repeated when transferring to thumbnail memory. Must perform a buffer read.

That is, when transferring to the thumbnail dedicated memory, the order of the transmission data based on the buffer map of FIG. 7 is {Y area 1-> Y area 2-> Y area 3-> Cb area-> Cb area-> Cr area-> Cr area. }.

In this case, in the case of the last slice, the Y region 4 overlaps the values of the Y region 3 so that the Y region 3 can be transferred once more to the thumbnail dedicated memory without actually being stored in the slice buffer. Determination of the last slice may be performed by parsing the slice vertical position during MPEG video decoding. Finally, the size of a thumbnail image generated from a 720x480 SD video is 180x120.

Therefore, when generating a thumbnail image for SD video, the interpolated pixel value is immediately calculated by averaging using two or four DC values at the moment when the DC value is input and stored in the slice buffer. It is possible to generate thumbnail images each enlarged 2 times in the horizontal and vertical directions without resizing the entire image through post-processing.

As described above, the present invention generates a thumbnail image using the DC value of the I picture of the MPEG-encoded video stream, but based on the size of the HD-quality video stream, the original image for the HD and SD video, respectively. Create thumbnail images with sizes of 1/64 and 1/16.

The present invention is applied to a program guide, storyboard, content summary, and other applications related to implicit information delivery of a video stream in a personal video recorder or a video retrieval device of a digital TV, regardless of the format of the video stream to the user. It is possible to provide a thumbnail image of a consistent size.

As described above, according to the apparatus and method for generating a thumbnail image according to the present invention, one thumbnail pixel is extracted for each subblock in the case of HD video, and four thumbnail pixels per subblock through interpolation in the case of SD video. By generating the, it is possible to solve the problem of structural thumbnail image size mismatch caused by the difference in the format of the stream to generate a thumbnail image of a consistent size. In addition, a relatively large thumbnail image may be generated while preventing excessive resolution loss in the thumbnail generation process.

Also, in case of SD video, the size of thumbnail video for SD video is generated by generating resized thumbnail data by averaging operation in memory storage process in slice unit without resizing the entire video after generating thumbnail video. It is effective to enlarge 4 times. In order to reduce the size of the memory required when performing the resizing operation, the DC value of the I picture is divided into predetermined units without performing a resizing operation on the entire thumbnail image to perform independent resizing for each unit. In this process, the aspect ratio of the original image can be maintained by applying the same scaling factor in the horizontal and vertical directions.

Those skilled in the art will appreciate that various changes and modifications can be made without departing from the spirit of the present invention.

Therefore, the technical scope of the present invention should not be limited to the contents described in the embodiments, but should be defined by the claims.

1 is a block diagram showing an embodiment of a thumbnail image generating apparatus according to the present invention

2 is a flowchart showing an embodiment of generating a thumbnail image for HD-quality video in the thumbnail image generating apparatus according to the present invention;

3 is a view showing an embodiment of a slice buffer for HD-quality video of the thumbnail image generating apparatus according to the present invention

4 is a flowchart illustrating an embodiment of generating a thumbnail image for SD video of a thumbnail image generating apparatus according to the present invention;

5 is a general conceptual diagram of a 2x magnification process of an image by linear interpolation;

FIG. 6 is a diagram illustrating an embodiment of a resizing process for thumbnail data of SD video of a thumbnail image generating apparatus according to the present invention; FIG.

7 is a view showing an embodiment of a slice buffer for the SD video of the thumbnail image generating apparatus according to the present invention

Explanation of symbols for main parts of the drawings

101: DC decoder 102: buffer controller

103: slice buffer 104: thumbnail memory interface

Claims (13)

A DC decoder for parsing only a DC value of an I picture from an input video stream; A slice buffer for storing the DC value output from the DC decoder in units of slices; And If the image including the DC value output from the DC decoder is HD level, the DC value is stored in the slice buffer as it is. If the SD level, the DC value of sub blocks included in each macro block of the image is interpolated. And a buffer controller which controls to enlarge the image in the slice buffer, and outputs the DC value stored in the slice buffer to a thumbnail dedicated memory to generate a thumbnail image. The method of claim 1, wherein the buffer controller If the input image is HD, the DC values for the four luminance blocks included in each macro block of the image and the DC values for the Cb and Cr blocks are controlled to be stored in the slice buffer in units of slices. Thumbnail image generating device. The method of claim 1, wherein the buffer controller If the input image is SD, the DC values of the four luminance blocks included in each macro block of the image are doubled in the horizontal and vertical directions by linear interpolation, respectively, and the DC values for the Cb and Cr blocks are 2. The thumbnail image generating apparatus of claim 2, wherein the thumbnail image is enlarged in the horizontal direction by linear interpolation and then stored in the slice buffer in units of slices. The method of claim 3, wherein the buffer controller When DC values for Cb and Cr blocks stored in the slice buffer are transferred to a thumbnail dedicated memory, DC values for Cb and Cr blocks corresponding to one thumbnail image line from the slice buffer are controlled by read point control of the slice buffer. Thumbnail image generating apparatus characterized by transmitting in duplicate. delete The method of claim 1, wherein the slice buffer And two buffers, wherein the two buffers perform double buffering under the control of the buffer controller. The method of claim 1, wherein the slice buffer The value of the luminance area in which the interpolated values are stored using only the DC value of the current slice under the control of the buffer controller is transferred to the thumbnail dedicated memory at the end of the current slice, and is interpolated using the DC values of the current slice and the next slice. And a value of the luminance region in which the image is stored is transmitted to the thumbnail dedicated memory at the end of the next slice. The method of claim 7, wherein the buffer controller In the case of the last slice, the value of the last luminance region is not stored in the slice buffer, and the thumbnail is characterized by overlapping and transmitting the value of the immediately previous luminance region by a read point control of the slice buffer when transferring to a thumbnail dedicated memory. Video generation device. In the method for generating a thumbnail image using a slice buffer for storing the input DC value, (a) parsing only a DC value of an I picture from an input video stream; (b) if the image including the DC value input in the parsing step is HD, storing the DC values for all sub-blocks included in each macro block of the image in the slice buffer in units of slices; (c) If the image including the DC value input in the parsing step is SD, the DC value of the subblocks included in each macroblock of the image is enlarged by 2 times in the horizontal and vertical directions through interpolation, and then sliced. Storing in the slice buffer; And and (d) transferring the DC values stored in the slice buffer to a thumbnail dedicated memory when the slice including the macro block ends. The method of claim 9, wherein step (c) If the input image is SD, the DC values for the four luminance blocks included in each macro block of the image are doubled in the horizontal and vertical directions through linear interpolation, and the DC values for the Cb and Cr blocks are linear. The method of generating a thumbnail image, characterized in that the image is enlarged only 2 times in the horizontal direction through interpolation and stored in the slice buffer. The method of claim 10, wherein step (d) When DC values for Cb and Cr blocks stored in the slice buffer are transferred to a thumbnail dedicated memory, DC values for Cb and Cr blocks corresponding to one thumbnail image line from the slice buffer are controlled by read point control of the slice buffer. The thumbnail image generating method characterized in that the transmission in duplicate. The method of claim 9, wherein step (d) The value of the luminance region in which the interpolated values are stored using only the DC value of the current slice is transmitted to the thumbnail dedicated memory at the end of the current slice, and the value of the luminance region in which the interpolated values are stored using the DC values of the current slice and the next slice is And transmitting the thumbnail image to the thumbnail dedicated memory at the end of the next slice. The method of claim 12, wherein step (d) In the case of the last slice, the value of the last luminance region is not stored in the slice buffer, and when transmitting to a thumbnail-only memory, the thumbnail of the previous luminance region is overlapped by the read point control of the slice buffer. How to create an image.