Video transcoding using syntactic and semantic clues
First Claim
Patent Images
1. A method for transcoding a compressed video, comprising the steps of:
- partitioning the compressed video into a plurality of hierarchical levels;
extracting features from each of the plurality of hierarchical levels;
selecting one of a plurality of conversion modes of a transcoder dependent on the features extracted from each of the plurality of hierarchical levels; and
transcoding the compressed video according to the selected conversion mode.
2 Assignments
0 Petitions
Accused Products
Abstract
A method for transcoding a compressed video partitions the compressed video into hierarchical levels, and extracts features from each of the hierarchical levels. One of a number of conversion modes of a transcoder is selected dependent on the features extracted from the hierarchical levels. The compressed video is then transcoded according to the selected conversion mode.
130 Citations
26 Claims
-
1. A method for transcoding a compressed video, comprising the steps of:
-
partitioning the compressed video into a plurality of hierarchical levels;
extracting features from each of the plurality of hierarchical levels;
selecting one of a plurality of conversion modes of a transcoder dependent on the features extracted from each of the plurality of hierarchical levels; and
transcoding the compressed video according to the selected conversion mode. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26)
mapping the low-level meta-data to rate-quality characteristics depending on content of the video.
-
-
6. The method of claim 5 wherein the rate-quality characteristics are partitioned into a plurality of classes.
-
7. The method of claim 6 including a first class where spatial quality is less than temporal quality, a second class where spatial quality is greater than temporal quality, and a third class where spatial quality is equal to temporal quality.
-
8. The method of claim 7 wherein content of the first class is transcoded with a discrete-summary transcoder, content of the second class is transcoded with a continuous-conversion transcoder at a relatively low frame rate, and content of the third class is transcoded with the continuous-conversion transcoder at a relatively high frame rate.
-
9. The method of claim 8 wherein the relatively low frame rate is less than ten frames per second, and the relatively high frame rate is greater than twenty frames per second.
-
10. The method of claim 6 including a first class where content is easy to compress, a second class where content is moderately difficult to compress, and a third class where content is difficult to compress.
-
11. The method of claim 10 wherein content of the first class is transcoded with a continuous-conversion transcoder at a relatively high frame rate, content of the second class is transcoded with a continuous-conversion transcoder at a moderate frame rate, and content of the third class is transcoded with the continuous-conversion transcoder at a relatively low high frame rate.
-
12. The method of claim 11 where the relatively high frame is greater than twenty frames per second, the moderate frame rate is between twenty and ten frames per second, the relatively low fame rate is less than ten frames per second.
-
13. The method of claim 5 wherein the low-level meta-data is based on a priority of content of the video.
-
14. The method of claim 5 wherein the low-level meta-data is based on a shape parameters of content of the video.
-
15. The method of claim 5 wherein the low-level meta-data is based on a texture of content of the video.
-
16. The method of claim 5 wherein the low-level meta-data is based on motion activity of content of the video.
-
17. The method of claim 5 wherein the low-level meta-data is based on a DCT coefficients of content of the video.
-
18. The method of claim 5 wherein high-level meta-data are encoded from low-level meta-data.
-
19. The method of claim 6 wherein the plurality of classes are ranked on a numeric scale.
-
20. The method of claim 1 wherein the transcoding produces a CBR bit stream.
-
21. The method of claim 1 wherein the transcoding produces a VBR bit stream.
-
22. The method of claim 1 further comprising steps of:
-
generating a descriptor scheme from the low-level meta-data; and
generating a variation of the compressed input bitstream according to the descriptor scheme.
-
-
23. The method of claim 1 wherein the transcoding is based on features extracted from a selected one of the hierarchical levels.
-
24. The method claim 23 wherein the features depend on spatial importance.
-
25. The method claim 23 wherein the features depend on temporal importance.
-
26. The method claim 23 wherein the features depend on compression difficulty.
Specification