Method and device for picture compression using a plurality of sampling patterns
View Patent ↗In order to realize effective picture compression without significantly impairing the resolution, it is proposed that the picture to be compressed be subdivided into a plurality of picture areas, for example into picture blocks each having 16×16 pixels, and the picture areas be subdivided into a plurality of layers, in particular into three layers, which correspond to different sampling patterns. The pixels (13-15) of a layer are in each case predicted from adjacent pixels of the other layers, in particular of the already previously processed and coded layers, the prediction area being coded. The invention is suitable in particular for use in video coders or video decoders with motion-compensated prediction, in order to compress the reference picture to be stored in a reference picture memory ( 5 ).
1. Method for compressing a picture area having a plurality of pixels described by corresponding pixel values, comprising the steps of:
subdividing the picture area to be compressed into at least a first sampling pattern and a second sampling pattern, the first and second sampling patterns being subsampled from the picture area, wherein the first and second sampling patterns include different pixels of the picture area to be compressed;
coding the pixels of the first sampling pattern;
after the coding step, determining a predicted pixel value and coding each pixel of the second sampling pattern from pixel values of the coded pixels of the first sampling pattern that are adjacent to the respective pixel of the second sampling pattern in the picture area; and
outputting a difference value between the predicted pixel value and an actual pixel value of the respective pixel of the second sampling pattern.
2. Method according to claim 1 , characterized in that a predicted pixel value is determined for the pixels of the first sampling pattern from pixels of the first sampling pattern that are adjacent in the picture area, and the difference value between the predicted pixel value and the actual pixel value of the respective pixel is output for further processing.
3. Method according to claim 2 , characterized in that the difference values determined for the pixels of the first sampling pattern are output in unquantized form for further processing.
4. Method according to claim 2 , characterized in that the difference values determined for the pixels of the first and second sampling patterns are coded.
5. Method according to claim 2 , characterized
in that the picture to be compressed is subdivided into the first and the second and also a third sampling pattern, the first, second and the third sampling pattern corresponding to different sampling patterns of the picture area,
in that a predicted pixel value is determined for each pixel of the third sampling pattern from pixel values of pixels of the first and second sampling patterns that are adjacent to the respective pixel in the picture area, and is output for further processing, and
in that the predicted pixel value is determined for each pixel of the third sampling pattern from the difference values which have been determined for pixels of the first and second sampling patterns that are adjacent to the respective pixel in the picture area, and is output for further processing.
6. Method according to claim 1 , characterized in that the first sampling pattern is obtained by subsampling, with a factor ¼, the picture area to be compressed.
7. Method according to claim 1 , characterized in that the second sampling pattern is obtained by subsampling, with a factor ¼, the picture area to be compressed.
8. Method according to claim 1 , characterized in that the difference values determined for the pixels of the second sampling pattern are quantized before they are output for further processing.
9. Method according to claim 1 , characterized
in that the picture to be compressed is subdivided into the first and the second and also a third sampling pattern, the first, second and the third sampling patterns corresponding to different sampling patterns of the picture area, and
in that a predicted pixel value is determined for each pixel of the third sampling pattern from pixel values of pixels of the first and second sampling patterns that are adjacent to the respective pixel in the picture area, and is output for further processing.
10. Method according to claim 1 , characterized
in that the picture area to be compressed is subdivided into a plurality of sampling patterns which correspond to different sampling patterns of the picture area and form a specific sampling pattern sequence, and
in that a predicted pixel value is determined for each pixel of a sampling pattern from pixel values of pixels of preceding sampling patterns in the sampling pattern sequence, which pixels are adjacent to the respective pixel in the picture area, and the difference value between the predicted pixel value and the actual pixel value of the respective pixel is output for further processing.
11. Method according to claim 10 , characterized in that, for each sampling pattern, a quantization of the difference values determined for the corresponding pixels of said sampling pattern is carried out, the quantization in each case being coarser than that for the preceding sampling patterns in the sampling pattern sequence.
12. The method as claimed in claim 10 , characterized in that the difference value determined for a pixel is output in coded form for further processing.
13. Method according to claim 12 , characterized in that the coding carried out is a Huffman coding.
14. Method according to claim 1 , characterized in that, in order to determine the predicted pixel value of a pixel, in each case two or four pixels arranged adjacent to the respective pixel in the picture area to be compressed are used.
15. Method according to claim 1 , characterized in that the predicted pixel value is determined by forming the mean value or median value of the corresponding adjacent pixels.
16. Method according to claim 1 , characterized in that the picture area used is in each case a picture area having 16×16 pixels.
17. Method according to claim 1 , characterized in that the picture area used is in each case a group of a plurality of picture blocks arranged next to one another in a picture line of a picture to be compressed.
18. Method according to claim 17 , characterized in that the group of picture blocks arranged next to one another corresponds to one quarter to one whole picture line of the picture to be compressed.
19. Method according to claim 1 , characterized in that the picture area to be compressed is processed in a progressive manner in accordance with the picture compression method.
20. Method according to claim 19 , characterized
in that the picture area to be compressed is processed in an interlaced manner in accordance with the picture compression, and
in that a changeover is made between progressive processing and interlaced processing in a manner dependent on the picture source of the picture area to be compressed.
21. Method according to claim 19 , characterized
in that the picture area to be compressed is processed in an interlaced manner in accordance with the picture compression, and
in that both progressive processing and interlaced processing are carried out for the picture area to be compressed and the more effective type of processing is subsequently selected.
22. Method according to claim 1 , characterized in that the picture area to be compressed is processed in an interlaced manner in accordance with the picture compression.
23. Device for picture compression, comprising:
a digital picture signal which describes corresponding pixel values of a picture area that is to be compressed and includes a plurality of pixels,
wherein the digital picture signal is fed to subsampling means for subdividing the picture area to be compressed into at least a first sampling pattern and a second sampling pattern, the first and second sampling patterns being subsampled from the picture area, wherein the first and second sampling patterns include different pixels of the picture area to be compressed;
means for coding the pixels of the first sampling pattern;
prediction means for determining a predicted pixel value and coding each pixel of the second sampling pattern from pixel values of the coded pixels of the first sampling pattern that are adjacent to the respective pixel of the second sampling pattern in the picture area to be compressed, and for outputting a difference value between the predicted pixel value and an actual pixel value of the respective pixel of the second sampling pattern; and
means for coding pixels of at least one additional sampling pattern based on the coded pixels of preceding sampling patterns and the outputted difference values.
24. Device according to claim 23 , further comprising means for coding the difference values output by the prediction means.
25. Device according to claim 24 , wherein the means for coding the difference values are configured for carrying out a Huffman coding of the difference values output by the prediction means.
26. Device according to claim 23 , characterized in that quantization means are provided for quantizing the difference values output by the prediction means.