Device for coding/decoding image information

• US 5,987,184 A
• Filed: 09/30/1997
• Issued: 11/16/1999
• Est. Priority Date: 09/30/1996
• Status: Expired due to Term

• Current Assignee: Pantech Incorporated
• Original Assignee: SK Hynix Inc.
• Inventors: Moon, Joo-Hee, Kim, Hae Kwang, Park, Chel-Soo, Kweon, Ji-Heon
• Primary Examiner(s): Tadayon, Bijan
• Application Number:
  US08/941,123
• Time in Patent Office: 777 Days
• Field of Search: 382/232, 382/233, 382/234, 382/235, 382/236, 382/238, 382/239, 382/240, 382/241, 382/242, 382/243, 382/244, 382/245, 382/246, 382/248, 382/250, 382/251, 382/252, 382/253, 341/67, 348/404, 348/384, 348/394, 348/395, 348/416, 348/700, 348/405, 358/432, 358/433
• US Class Current: 382/250
• CPC Class Codes: H04N 19/20 using video object coding
• CPC Class Codes: H04N 19/593 involving spatial prediction techniques
• CPC Class Codes: H04N 19/61 in combination with predictive coding

Abstract

A device for coding/decoding image information is provided to perform a boundary block merge of macro blocks of image information. Further, the DCT coefficients of a block are copied, if that is converted to an exterior block from a boundary block out of left-sided and upper-sided blocks for the purpose of performing a prediction of the DCT coefficients of the block'"'"'s sub blocks. Thus, the efficiency of a coding/decoding process is enhanced.

Claims

1. A device for coding information, comprising:

a motion estimating section for receiving an object image signal to estimate motion in the units of a macro block, said object image signal comprising a plurality of macro blocks, said macro blocks comprising a plurality of pixels, said motion estimation section outputting motion information of a macro block;

a motion compensating section, said motion compensating section receiving said motion information, said motion compensating section compensating the motion information of the macro block, wherein said motion compensating section outputs a motion compensated macro block unit;

a subtractor, said subtractor receiving said object image signal and said motion compensated macro block, said subtractor detecting the difference between said motion compensated macro block and said object, image signal;

an image signal coding section, said object image signal coding section receiving the output of the subtractor, said image coding section dividing said macro blocks into a plurality of interior sub blocks, exterior sub blocks, and boundary sub blocks, said object image signal coding section comparing pixel values corresponding to respective blocks so as to determine whether a block received is a boundary block, and performing a boundary block merge;

an adder, said adder receiving the output of said image coding section, and said motion compensated macro block unit, said adder detecting the sum of the output of said image coding section and said motion compensated macro block unit;

a previous image detecting section for detecting the previous image signal from the output of the adder, and generating it to the motion estimating section and the motion compensating section;

a multiplexer, said multiplexer receiving the output of said image coding section, and said motion information of said macro block, said multiplexer for multiplexing the output of said image of said image signal coding section and said motion information of said macro block; and

a buffer, said buffer receiving the output of said multiplexer, and transmitting said output of said multiplexer.

2. The device as defined in claim 1, wherein the image signal coding section comprises:

a boundary block merging section for receiving the object image signal, comparing pixel values corresponding to respective blocks so as to determine whether the block received is a boundary block, and performing a boundary block merge;

said boundary block merging section outputting a merged image signal;

a discrete cosine transformer, said discrete cosine transformer receiving said merged image signal, said discrete cosine transformer for performing a discrete cosine transformation of said merged image signal, said discrete cosine transformer outputting a transformed coefficient signal;

a quantizing section, said quantizing section receiving said transformed coefficient signal, said quantizing section for quantizing the transformed coefficient signal received from the discrete cosine transformer, said quantizing section outputting a quantized transformed coefficient signal;

a transformed coefficient and coded block pattern coding section for receiving the quantized transformed coefficient signal from the quantizing section and the merged image signal from the boundary block merging section, said transformed coefficient and coded block pattern comparing pixel values corresponding to the respective blocks, and selectively generating texture information;

an inverse quantizing section, said inverse quantizing section receiving said quantized transformed coefficient signal, said inverse quantizing section inverse-quantizing the quantized transformed coefficient signal and extracting the transformed coefficient signal;

an inverse discrete cosine transformer, said inverse discrete cosine transformer receiving said extracted coefficient signal, said inverse discrete cosine transformer performing an inverse discrete cosine transformation of the extracted transformed coefficient; and

a boundary block split section for receiving the image signal from the inverse discrete cosine transformer and the merged information from the boundary block merging section, and generating decoded data.

3. The device as defined in claim 1, wherein the image signal coding section comprises:

a macro block scanning section for scanning the macro blocks in the units of a sub block;

a sub block merging section, said sub block merging section comprising a plurality of requirements for merging sub blocks, said sub block merging section merging the sub blocks scanned by the macro block scanning section when the sub blocks meet said requirements for a merge; and

a texture-based image coding section for coding the blocks merged by the sub block merging section.

4. The device as defined in claim 1, wherein the image coding section comprises means for coding the object image signal by performing a merge over a frame image without restriction on the macro block, said image signal coding section further comprising:

a frame image merging section for merging frame images; and

a texture-based image coding section for coding the image signal merged by the frame image merging section.

5. The device as defined in claim 3, wherein the macro block scanning section stores profile information of an object image by dividing the macro block into blocks 1 to 4 and scanning respective sub blocks, and the sub block merging section searches for boundary blocks and compares the profile information of coordinates corresponding to respective boundary blocks, performing a merge.

6. The device as defined in claim 3, wherein if there is no pixel overlapped between a first boundary block rotated in a 180-degree arc and a second boundary block, said first boundary block is merged into said second boundary block, and coded.

7. A device for decoding image information, said image information comprising a plurality of macro blocks, said macro blocks comprising a plurality of interior sub blocks, exterior sub blocks, and boundary sub blocks, said macro blocks further comprising a plurality of pixels, said device comprising:

a demultiplexer for demultiplexing image and motion information transferred from a coding section;

a motion decoding section for detecting the motion information from the output of the demultiplexer;

an image signal decoding section for receiving the image information from the demultiplexer, comparing pixel values corresponding to respective blocks, and performing a boundary block division and a decoding, said image signal decoding section outputting image and motion information;

a motion compensating section for receiving the image and motion information detected by the motion decoding section, said motion compensating section compensating said motion information for outputting motion compensated information;

an image reconstructing section for receiving the motion-compensated information from the motion compensating section, and reconstructing an image; and

an image memory for storing and generating the image reconstructed by the image reconstructing section.

8. The device as defined in claim 7, wherein the image signal decoding section comprises:

a transformed coefficient and coded block pattern decoding section for receiving the image information from the demultiplexer, comparing pixel values corresponding to respective blocks so as to determine whether the block received is a boundary block, and performing a boundary block division and a prediction to decode the image information;

an inverse quantizing section for multiplying the image information, decoded by the transformed coefficient and coded block pattern decoding section, by a quantized coefficient, said inverse quantizing section outputting a transformed coefficient;

an inverse discrete cosine transformer for performing an inverse discrete cosine transformation of the transformed coefficient inverse-quantized by the inverse quantizing section, and generating said transformed coefficient as image data in a space region; and

a boundary block split section for receiving the image signal from the inverse discrete cosine transformer, and generating image data in the units of a macro block.

9. A device for coding image information, which is an image encoder for splitting an image into object and background images that have predetermined shape information, said device comprising:

a Video Object Plane (VOP) definition section for receiving an image signal, and defining a VOP;

a motion estimating section for receiving the image signal in macro block units from the VOP definition section, said said macro blocks comprising a plurality of pixels, said motion estimating section estimating a motion and outputting motion information;

a motion compensating section for receiving said motion information from the motion estimating section, and compensating the motion information, said motion compensating section outputting a motion compensated VOP;

a subtractor for detecting the difference between the motion-compensated VOP from the motion compensating section and the defined VOP from the VOP definition section;

an image signal coding section for receiving the output of the subtractor and the predetermined shape information, said image signal coding section dividing said macro blocks into interior sub blocks, exterior sub blocks and boundary blocks, merging macro blocks by performing a boundary block merge, and coding the merged macro blocks in the units of a sub block, said image signal coding section outputting interior information;

an adder for detecting the sum of the motion compensated VOP motion-compensated by the motion compensating section and interior information coded by the image signal coding section;

a previous VOP detecting section for detecting the previous VOP from the output of the adder, and generating it to the motion estimating section and the motion compensating section;

a shape coding section for receiving the image in macro block unit from the VOP definition section, and coding the shape information;

a multiplexer for multiplexing the motion information estimated by the motion estimating section, the interior information coded by the image signal coding section, and the shape information coded by the shape coding section; and

a buffer for transmitting the output of the multiplexer as a bit stream.

10. The device as defined in claim 9, wherein the output signal of the shape coding section is transferred into the motion estimating section, the motion compensating section, and the image signal coding section, said output signal being used in estimating the motion, compensating the motion, and coding the interior information.

11. The device as defined in claim 9, wherein the image signal coding section comprises:

a macro block scanning section for scanning the macro blocks in the units of a sub block;

a sub block merging section, said sub block merging section comprising a plurality of requirements for merging sub blocks, said sub block merging section merging the sub blocks scanned by the macro block scanning section when the sub blocks meet said requirements for a merge; and

a texture-based image coding section for coding the blocks merged by the sub block merging section.

12. The device as defined in claim 9, wherein the image signal coding section comprises means for coding the image signal by performing a merge over a frame image without restriction on a macro block said image coding section further comprising:

a frame image merging section for merging frame images, said frame image merging section outputting an image signal; and

a texture-based image coding section for coding the image signal merged by the frame image merging section.

13. The device as defined in claim 9, wherein said sub blocks further comprise a background part and an object part, and the image signal coding section comprises:

a boundary block padding section for receiving the image signal and the original shape information, and replacing the background part of the sub blocks with the mean value of the object part, said boundary block padding section outputting reconstructed shape information;

a boundary block merging section for receiving the image signal and the reconstructed shape information from the boundary block padding section, and performing a boundary block merge, said boundary block merging section outputting a merged image signal;

a discrete cosine transformer for performing a discrete cosine transformation of the merged image signal received from the boundary block merging section, said discrete cosine transformer outputting a transformed coefficient, wherein said transformed coefficient comprises DC and AC coefficients;

a quantizing section for quantizing the transformed coefficient received from the discrete cosine transformer, said quantizing section outputting a quantized transformed coefficient;

a transformed coefficient and coded block pattern coding section for receiving the quantized transformed coefficient quantized by the quantizing section and merged image signal transferred from the boundary block merging section, and selectively coding texture information;

an inverse quantizing section for inverse-quantizing the quantized transformed coefficient quantized by the quantizing section, and extracting a transformed coefficient;

an inverse discrete cosine transformer for performing an inverse discrete cosine transformation of the transformed coefficient received from the inverse quantizing section; and

a boundary block split section for receiving the image signal from the inverse discrete cosine transformer and the reconstructed shape information, and generating decoded data.

14. The device as defined in claim 13, wherein when the transformed coefficient and coded block pattern coding section generates texture information and the merged image signal generated by the boundary block merging section indicates that a current sub block is an exterior sub block converted from a boundary sub block by a boundary block merge, the quantized DC and AC coefficients of the merged blocks are copied into the current sub block.

15. The device as defined in claim 13, wherein when the transformed coefficient and coded block pattern coding section generates texture information, and the merged image signal indicates that a current sub block is an exterior sub block converted from a boundary sub block by a boundary block merge, the DC coefficient is set to 2.sup.(bits per pixel-1) where bits per pixel is the number of bits indicating the pulse code modulation data of one pixel, and the AC coefficient is set to 0.

16. The device as defined in claim 13, wherein when the transformed coefficient and coded block pattern coding section generates texture information, and the merged image signal indicates that a current sub block is an exterior sub block converted from a boundary sub block by a boundary block merge, a prediction is performed by copying the AC and DC coefficients of a first block, wherein said first block is not an exterior block, out of left-sided and upper-sided blocks, into the current sub block.

17. The device as defined in claim 13, wherein when the transformed coefficient and coded block pattern coding section generates texture information, and the merged image signal indicates that a current sub block is an exterior sub block converted from a boundary sub block by a boundary block merge, a prediction is performed by copying the quantized DC coefficient of the merged block into the current sub block and setting the AC coefficient to 0.

18. The device as defined in claim 13, wherein when the transformed coefficient and coded block pattern coding section generates texture information, and the merged image signal indicates that a current sub block is an exterior sub block converted from a boundary sub block by a boundary block merge, a prediction is performed by copying the quantized AC and DC coefficients of the merged block into the current sub block.

19. A device for decoding image information, wherein said image information comprises video object plane (VOP) data, a plurality of macro blocks, said macro blocks comprising a plurality of interior sub blocks, exterior sub blocks, and boundary sub blocks, said macro blocks further comprising a plurality of pixels, said device being an image decoder for splitting a coded image into object and background images that have predetermined shape information and comprising:

a demultiplexer for demultiplexing the image information received, said demultiplexer outputting image information;

a shape decoding section for decoding the image information transferred from the demultiplexer, said shape decoding section outputting shape information;

a motion decoding section for detecting motion information from the output of the demultiplexer;

an image signal decoding section for decoding the image information transferred from the demultiplexer to generate a decoded image signal;

a motion compensating section for receiving the motion information detected by the motion decoding section, the decoded shape information decoded by the shape decoding section, and the VOP data so as to compensate a motion, said motion-compensating section outputting motion compensated information; and

a VOP reconstructing section for receiving the motion information decoded by the motion decoding section, the motion-compensated information from the motion compensating section, and the decoded image signal decoded by the image signal decoding section, dividing merged blocks from said signals, performing a prediction, and reconstructing a VOP.

20. The device as defined in claim 19, wherein the VOP reconstructing section comprises:

a merged information extracting section for receiving the shape information decoded by the shape decoding section, and extracting merged information by a boundary block merge;

a transformed coefficient and coded block pattern decoding section for receiving the merged information from the merged information extracting section, the motion-compensated information from the motion compensating section, and the image signal from the decoded image signal decoding section, and decoding the coded image information, said transformed coefficient and coded block pattern decoding section outputting image information, and selectively generating texture information;

an inverse quantizing section for multiplying the image information, decoded by the transformed coefficient and coded block pattern decoding section, by the quantized coefficient, said inverse quantizing section outputting a transformed coefficient, wherein said transformed coefficient comprises both DC and AC coefficients;

an inverse discrete cosine transformer for performing an inverse discrete cosine transformation of the transformed coefficient inverse-quantized by the inverse quantizing section, and generating said transformed coefficient as an image signal in a space region; and

a boundary block split section for receiving the image signal from the inverse discrete cosine transformer and the shape information from the shape decoding section, and generating decoded VOP data in the units of a macro block.

21. The device as defined in claim 19, wherein if the transformed coefficient and coded block pattern coding section generates texture information, and the merged information indicates that the current sub block is an exterior sub block converted from a boundary sub block by a boundary block merge, the DC coefficient is set to 2.sup.(bits per pixel-1 where bits per pixel is the number of bits indicating the pulse code modulation data of one pixel, and the AC coefficient is set to 0.

22. The device as defined in claim 19, wherein if the transformed coefficient and coded block pattern coding section generates texture information, and the merged information indicates that the current sub block is an exterior sub block converted from a boundary sub block by a boundary block merge, a prediction is performed by copying the AC and DC coefficients of a block, which is not an exterior block out of left-sided and upper-sided blocks, into the current sub block.

23. The device as defined in claim 19, wherein when the transformed coefficient and coded block pattern coding section generates texture information, and the merged information indicates that a current sub block is an exterior sub block converted from a boundary sub block by a boundary block merge, a prediction is performed by copying the quantized DC coefficient of merged block into the current sub block and setting the AC coefficient to 0.

24. The device as defined in claim 19, wherein if the transformed coefficient and coded block pattern coding section generates texture information, and the merged information indicates that the current sub block is an exterior sub block converted from a boundary sub block by a boundary block merge, a prediction is performed by copying the quantized AC and DC coefficients of the merged block into the current sub block.