Apparatus and method for recovery of lost/damaged data in a bitstream of data based on compatibility of adjacent blocks of data
First Claim
Patent Images
1. A method for receiving lost/damaged data in a bitstream of encoded data comprising the steps of:
- generating a plurality of candidate decodings for a group of related blocks within the bit stream;
generating at least one internal correlation measure for each block based upon at least an image portion of the candidate decodings to determine the intra-block correlation of adjacent pixels within the block;
determining an external measure compatibility between adjacent blocks of each block from the group of blocks to determine the inter-block compatibility;
combining the internal correlation measure of the block and the external compatibility measure of related blocks to generate a combined measure of the intra-block correlation and the inter-block compatibility; and
selecting a candidate decoding from the plurality of candidate decodings based upon the combined measure.
1 Assignment
0 Petitions
Accused Products
Abstract
A system and method for receiving damage data in a bitstream of encoded data is disclosed. A plurality of candidate decodings for a plurality of blocks are generated. A measure is generated for each block based upon at least an image portion of the candidate decodings. Compatibility is determined between adjacent blocks of the plurality of blocks. A candidate decoding is selected based upon the measure and the compatibility of adjacent blocks. In one embodiment, this is used in the transmission of video signals over a potentially lossy communications channel.
-
Citations
82 Claims
-
1. A method for receiving lost/damaged data in a bitstream of encoded data comprising the steps of:
-
generating a plurality of candidate decodings for a group of related blocks within the bit stream;
generating at least one internal correlation measure for each block based upon at least an image portion of the candidate decodings to determine the intra-block correlation of adjacent pixels within the block;
determining an external measure compatibility between adjacent blocks of each block from the group of blocks to determine the inter-block compatibility;
combining the internal correlation measure of the block and the external compatibility measure of related blocks to generate a combined measure of the intra-block correlation and the inter-block compatibility; and
selecting a candidate decoding from the plurality of candidate decodings based upon the combined measure. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8)
-
-
9. A method for receiving lost/damaged data in a bitstream of encoded data comprising the steps of:
-
generating a plurality of candidate decodings for a plurality of blocks;
measuring for each block an internal smoothness measure of at least an image portion of the candidate decodings;
determining an external compatibility between adjacent blocks of the plurality of blocks;
combining the internal smoothness measure and the external compatibility measure of adjacent blocks to generate a combined measure; and
selecting a candidate decoding from the plurality of candidate decodings based upon the combined measure. - View Dependent Claims (10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25)
normalizing x[i][j] that is defined as {tilde over (x)}[i][b] where x[j][j] is the decoded value at location [i,j]; and
computing smoothness using the x[i][j], Laplacian kernel L[m][n], and a block Laplacian value Lx, where Lx is determined according to the following;
-
-
12. The method as set forth in claim 11 where the {tilde over (x)}[i][j] is computed according to the following equation:
-
13. The method as set forth in claim 11, wherein K is 8.
-
14. The method as set forth in claim 9, wherein measuring smoothness further comprises:
-
normalizing q[i][j], wherein q[i][j] is the Q code value at location [i,j] in an encoded image and q′
[i][j] is the normalized value of q[i][j];
computing smoothness using q′
[i][j], Laplacian kernel L[m][n], and a block Laplacian value Lq, where Lq equals
-
-
15. The method as set forth in claim 14, wherein q′
- [i][j] is computed according to the following equation;
- [i][j] is computed according to the following equation;
-
16. The method as set forth in claim 14, wherein K is 8.
-
17. The method as set forth in claim 9, wherein the smoothness measure comprises a second order image surface property.
-
18. The method as set forth in claim 9, wherein the smoothness measure comprises an N order image surface property, where N>
- 1.
-
19. The method as set forth in claim 9, wherein measuring the smoothness further comprises:
-
determining if the image portion includes motion;
if the image portion includes motion, computing a smoothness measurement for a first field of the image portion and a second field of the image portion, and averaging the smoothness measurements together to generate the smoothness measure; and
if the image portion does not include motion, generating the smoothness measure for the image portion.
-
-
20. The method as set forth in claim 9, wherein determining further comprises evaluating spatial correlation to determine motion.
-
21. The method as set forth in claim 9, wherein determining a measure of compatibility between adjacent blocks further comprises determining the ratio between adjacent blocks and the total number of boundary pixels at the candidate block.
-
22. The method as set forth in claim 21, wherein a boundary pixel identified as compatible if it lies within a determined valid range of compatible boundary pixels.
-
23. The method as set forth in claim 9, further comprising generating a combined measure used to select a candidate decoding.
-
24. The method as set forth in claim 23, wherein the combined measure is determined according the following equation:
-
25. The method as set forth in claim 23, wherein the combined measure is determined according the following equation:
-
26. A processing system configured to receive lost/damaged data in a bitstream of encoded data comprising:
-
circuitry configured to generate a plurality of candidate decodings for a plurality of blocks;
circuitry configured to generate an internal correlation measure for each block based upon at least an image portion of the candidate decodings;
circuitry configured to determine an external compatibility measure between adjacent blocks of the plurality of blocks;
circuitry configured to combine the internal correlation measure and the external compatibility measure of adjacent blocks to generate a combined measure; and
circuitry configured to select a candidate decoding from the plurality of candidate decodings based upon the combined measure. - View Dependent Claims (27, 28, 29, 30, 31, 32, 33)
-
-
34. A processing system for receiving lost/damaged data in a bitstream of encoded data comprising:
-
circuitry configured to generate a plurality of candidate decodings for a plurality of blocks;
circuitry configured to measure for each block the internal smoothness of at least an image portion of the candidate decodings;
circuitry configured to determine external compatibility measure between adjacent blocks of the plurality of blocks;
circuitry configured to combine the internal smoothness measure and the external compatibility measure of adjacent blocks to generate a combined measure; and
circuitry configured to select a candidate decoding from the plurality of candidate decodings based upon the combined measure. - View Dependent Claims (35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50)
normalizing x[i][j] that is defined as {tilde over (x)}[i][j] where x[i][j] is the decoded value at location [i,j]; and
computing smoothness using the x[i][j], Laplacian kernel L[m][n], and a block Laplacian value Lx, where Lx is determined according to the following;
-
-
37. The processing system as set forth in claim 36, where the {tilde over (x)}[i][j] is computed according to the following equation:
-
38. The processing system as set forth in claim 36, wherein K is 8.
-
39. The processing system as set forth in claim 34, wherein measuring smoothness further comprises:
-
normalizing q[i][j], wherein q[i][j] is the Q code value at location [i,j] in an encoded image and q′
[i][j] is the normalized value of q[i][j];
computing smoothness using q′
[i][j], Laplacian kernel L[m][n], and a block Laplacian value Lq, where Lq equals
-
-
40. The processing system as set forth in claim 39, wherein q′
- [i][j] is computed according to the following equation;
- [i][j] is computed according to the following equation;
-
41. The processing system as set forth in claim 39, wherein K is 8.
-
42. The processing system as set forth in claim 34, wherein the smoothness measure comprises a second order image surface property.
-
43. The processing system as set forth in claim 34, wherein the smoothness measure comprises an N order image surface property, where N>
- 1.
-
44. The processing system as set forth in claim 34, wherein measuring the smoothness further comprises:
-
determining if the image portion includes motion;
if the image portion includes motion, computing a smoothness measurement for a first field of the image portion and a second field of the image portion, and averaging the smoothness measurements together to generate the smoothness measure; and
if the image portion does not include motion, generating the smoothness measure for the image portion.
-
-
45. The processing system as set forth in claim 34, wherein determining further comprises evaluating spatial correlation to determine motion.
-
46. The processing system as set forth in claim 34, wherein determining a measure of compatibility between adjacent blocks further comprises determining the ratio between adjacent blocks and the total number of boundary pixels at the candidate block.
-
47. The processing system as set forth in claim 46, wherein a boundary pixel identified as compatible if it lies within a determined valid range of compatible boundary pixels.
-
48. The processing system as set forth in claim 34, further comprising generating a combined measure used to select a candidate decoding.
-
49. The processing system as set forth in claim 48, wherein the combined measure is determined according the following equation:
-
50. The processing system as set forth in claim 48, wherein the combined measure is determined according the following equation:
-
51. A computer readable medium containing executable instructions which, when executed in a processing system, cause the system to receive lost/damaged data in a bitstream of encoded data, the system performing the steps comprising:
-
generating a plurality of candidate decodings for a plurality of blocks;
generating an internal correlation measure for each block based upon at least an image portion of the candidate decodings;
determining an external compatibility measure between adjacent blocks of the plurality of blocks;
combining the internal correlation measure and the external compatibility measure of adjacent blocks to generate a combined measure; and
selecting a candidate decoding from the plurality of candidate decodings based upon the combined measure. - View Dependent Claims (52, 53, 54, 55, 56, 57, 58)
-
-
59. A computer readable medium containing executable instructions which, when executed in a processing system, cause the system to receive lost/damaged data in a bitstream of encoded data, the system performing the steps comprising:
-
generating a plurality of candidate decodings for a plurality of blocks;
measuring for each block the internal smoothness of at least an image portion of the candidate decodings;
determining an external compatibility measure between adjacent blocks of the plurality of blocks;
combining the internal smoothness measure and the external compatibility measure of adjacent blocks to generate a combined measure; and
selecting a candidate decoding from the plurality of candidate decodings based upon the combined measure. - View Dependent Claims (60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75)
normalizing x[i][j] that is defined as {tilde over (x)}[i][j] where x[i][j] is the decoded value at location [i,j]; and
computing smoothness using the x[i][j], Laplacian kernel L[m][n], and a block Laplacian value Lx, where Lx is determined according to the following;
-
-
62. The computer readable medium as set forth in claim 61, where the {tilde over (x)}[i][j] is computed according to the following equation:
-
63. The computer readable medium as set forth in claim 61, wherein K is 8.
-
64. The computer readable medium as set forth in claim 59, wherein the step of measuring smoothness comprises instructions which when executed perform the steps of:
-
normalizing q[i][j], wherein q[i][j] is the Q code value at location [i,j] in an encoded image and q′
[i][j] is the normalized value of q[i][j];
computing smoothness using q′
[i][j], Laplacian kernel L[m][n], and a block Laplacian value Lq, where Lq equals
-
-
65. The computer readable medium as set forth in claim 64, wherein q′
- [i][j] is computed according to the following equation;
- [i][j] is computed according to the following equation;
-
66. The computer readable medium as set forth in claim 64, wherein K is 8.
-
67. The computer readable medium as set forth in claim 59, wherein the smoothness measure comprises a second order image surface property.
-
68. The computer readable medium as set forth in claim 59, wherein the smoothness measure comprises an N order image surface property, where N>
- 1.
-
69. The computer readable medium as set forth in claim 59, wherein the step of measuring the smoothness comprises instructions which when executed perform the steps of:
-
determining if the image portion includes motion;
if the image portion includes motion;
computing a smoothness measurement for a first field of the image portion and a second field of the image portion; and
averaging the smoothness measurements together to generate the smoothness measure; and
if the image portion does not include motion, generating the smoothness measure for the image portion.
-
-
70. The computer readable medium as set forth in claim 59, wherein the step of determining comprises instructions which when executed perform the step of evaluating spatial correlation to determine motion.
-
71. The computer readable medium as set forth in claim 59, wherein the step of determining a measure of compatibility between adjacent blocks comprises instructions which when executed perform the step of determining the ratio between adjacent blocks and the total number of boundary pixels at the candidate block.
-
72. The computer readable medium as set forth in claim 71, wherein a boundary pixel identified as compatible if it lies within a determined valid range of compatible boundary pixels.
-
73. The computer readable medium as set forth in claim 59, further comprising instructions which when executed perform the step of generating a combined measure used to select a candidate decoding.
-
74. The computer readable medium as set forth in claim 73, wherein the combined measure is determined according the following equation:
-
75. The computer readable medium as set forth in claim 73, wherein the combined measure is determined according the following equation:
-
76. An apparatus for receiving lost/damaged data in a bitstream of encoded data comprising:
-
means for generating a plurality of candidate decodings for a plurality of blocks;
means for measuring for each block the internal smoothness of at least an image portion of the candidate decodings;
means for determining an external compatibility measure between adjacent blocks of the plurality of blocks;
means for combining the internal smoothness measure and the external compatibility measure of adjacent blocks to generate a combined measure; and
means for selecting a candidate decoding from the plurality of candidate decodings based upon the combined measure. - View Dependent Claims (77, 78, 79, 80, 81, 82)
means for normalizing x[i][j] that is defined as {tilde over (x)}[i][j] where x[i][j] is the decoded value at location [i,j]; and
means for computing smoothness using the x[i][j], Laplacian kernel L[m][n], and a block Laplacian value Lx, where Lx is determined according to the following;
-
-
79. The apparatus as set forth in claim 78, where the {tilde over (x)}[i][j] is computed according to the following equation:
-
80. The apparatus as set forth in claim 76, wherein the means for measuring smoothness comprises:
-
means for normalizing q[i][j], wherein q[i][j] is the Q code value at location [i,j] in an encoded image and q′
[i][j] is the normalized value of q[i][j];
means for computing smoothness using q′
[i][j], Laplacian kernel L[m][n], and a block Laplacian value Lq, where Lq equals
-
-
81. The apparatus as set forth in claim 80, wherein q′
- [i][j] is computed according to the following equation;
- [i][j] is computed according to the following equation;
-
82. The apparatus as set forth in claim 80, wherein K is 8.
Specification