Significance map encoding and decoding using partition selection
View Patent ↗Methods of encoding and decoding for video data are describe in which significance maps are encoded and decoded using non-spatially-uniform partitioning of the map into parts, wherein the bit positions within each part are associated with a given context. Example partition sets and processes for selecting from amongst predetermined partition sets and communicating the selection to the decoder are described.
1. A method of encoding a significance map for a transform unit using an image or video encoder, the method comprising:
for each bit position in the significance map with a non-zero coefficient,
determining a context for that bit position based upon a partition set,
encoding a bit value at that bit position based on the determined context, and
updating the context based on that bit value,
wherein the transform unit is sized 4×4;
wherein the partition set assigns contexts to bit positions in accordance with a block-based mapping given by:
0,
1,
2,
3,
4,
5,
2,
3,
6,
6,
7,
7,
8,
8,
7,
and wherein the above integers represent the contexts assigned to the bit positions of a 4×4 block significance map.
2. The method claimed in claim 1 , wherein determining includes selecting the partition set from among a plurality of partition sets based upon text type and transform unit size, and wherein the text type is luma.
3. The method claimed in claim 1 , further comprising selecting the partition set from amongst a plurality of partition sets based upon text type, transform unit size, and selection information, wherein the selection information designates one of the plurality of partition sets for use with a particular transform unit size and text type.
4. The method claimed in claim 3 , wherein the selection information is encoded in a slice header or sequence header, and wherein the selection information has a first part identifying one or more of the plurality of partition sets, and a second part assigning to each predefined combination of text type and transform unit size one of the identified partition sets.
5. The method claimed in claim 1 , further comprising selecting the partition set from amongst a plurality of partition sets, and further comprising determining for each transform unit whether a slice size of a slice being encoded to that point has exceeded a threshold value and, if so, switching from the partition set to a refined partition set for encoding of subsequent significance maps within the slice, wherein the partition set is a subset of the refined partition set.
6. The method claimed in claim 1 , wherein determining includes selecting the partition set from among a plurality of partition sets based upon text type and transform unit size, and wherein the text type is chroma and wherein an offset is used for context assignment.
7. An encoder for encoding a significance map for a transform unit, the decoder comprising:
a processor;
a memory; and
an encoding application stored in memory and containing instructions for configuring the processor to:
for each bit position in the significance map with a non-zero coefficient,
determine a context for that bit position based upon a partition set,
encoding a bit value at that bit position based on the determined context, and
update the context based on that reconstructed bit value;
wherein the transform unit is sized 4×4;
wherein the partition set assigns contexts to bit positions in accordance with a block-based mapping given by:
0,
1,
2,
3,
4,
5,
2,
3,
6,
6,
7,
7,
8,
8,
7,
and wherein the above integers represent the contexts assigned to the bit positions of a 4×4 block significance map.
8. The encoder claimed in claim 7 , wherein the processor is configured to select the partition set from among a plurality of partition sets based upon text type and transform unit size, and wherein the text type is luma.
9. The encoder claimed in claim 7 , wherein the processor is further configured to select the partition set from amongst a plurality of partition sets based upon text type, transform unit size, and selection information, and wherein the selection information designates one of the plurality of partition sets for use with a particular transform unit size and text type.
10. The encoder claimed in claim 9 , wherein the selection information is encoded in a slice header or sequence header, and wherein the selection information has a first part identifying one or more of the plurality of partition sets, and a second part assigning to each predefined combination of text type and transform unit size one of the identified partition sets.
11. The encoder claimed in claim 7 , wherein the processor is further configured to select the partition set from amongst a plurality of partition sets, and the processor is further configured to determine for each transform unit whether a slice size of a slice being encoded to that point has exceeded a threshold value and, if so, to switch from the partition set to a refined partition set for encoding of subsequent significance maps within the slice, wherein the partition set is a subset of the refined partition set.
12. The encoder claimed in claim 7 , wherein the processor is configured to select the partition set from among a plurality of partition sets based upon text type and transform unit size, and wherein the text type is chroma and wherein an offset is used for context assignment.
13. A non-transitory processor-readable medium storing processor-executable instructions which, when executed, configures one or more processors to perform a method of encoding a significance map for a transform unit, the method comprising:
for each bit position in the significance map with a non-zero coefficient,
determining a context for that bit position based upon a partition set,
encoding a bit value at that bit position based on the determined context, and
updating the context based on that encoded bit value,
wherein the transform unit is sized 4×4;
wherein the partition set assigns contexts to bit positions in accordance with a block-based mapping given by:
0,
1,
2,
3,
4,
5,
2,
3,
6,
6,
7,
7,
8,
8,
7,
and wherein the above integers represent the contexts assigned to the bit positions of a 4×4 block significance map.
14. The non-transitory processor-readable medium claimed in claim 13 , wherein determining includes selecting the partition set from among a plurality of partition sets based upon text type and transform unit size, and wherein the text type is luma.
15. The non-transitory processor-readable medium claimed in claim 13 , wherein determining includes selecting the partition set from among a plurality of partition sets based upon text type and transform unit size, and wherein the text type is chroma and wherein an offset is used for context assignment.
16. The non-transitory processor-readable medium claimed in claim 13 , the method further comprising selecting the partition set from amongst a plurality of partition sets based upon text type, transform unit size, and selection information, wherein the selection information designates one of the plurality of partition sets for use with a particular transform unit size and text type.
17. The non-transitory processor-readable medium claimed in claim 16 , wherein the selection information is encoded in a slice header or sequence header, and wherein the selection information has a first part identifying one or more of the plurality of partition sets, and a second part assigning to each predefined combination of text type and transform unit size one of the identified partition sets.
18. The non-transitory processor-readable medium claimed in claim 13 , the method further comprising selecting the partition set from amongst a plurality of partition sets, and further comprising determining for each transform unit whether a slice size of a slice being encoded to that point has exceeded a threshold value and, if so, switching from the partition set to a refined partition set for encoding of subsequent significance maps within the slice, wherein the partition set is a subset of the refined partition set.