Granted Patent B2

US 12,526,455 · App. 18/818,310 · Granted Jan 13, 2026

Inventors: Santiago De Luxán Hernández (Berlin, DE); Valeri George (Berlin, DE); Gayathri Venugopal (Berlin, DE); Jens Brandenburg (Berlin, DE); Benjamin Bross (Berlin, DE); Heiko Schwarz (Berlin, DE); Detlev Marpe (Berlin, DE); Thomas Wiegand (Berlin, DE)

Assignee: Fraunhofer-Gesellschaft Zur Foerderung der angewandten Forschung e.V.

H04N19/61H04N19/159H04N19/176H04N19/46H04N19/625H04N19/70

Abstract

Claims (50)

1 . A video decoder, the video decoder configured to:

decode, from a data stream, an intra-prediction mode for an intra-predicted block of a picture;

decode, from the data stream, information on a partitioning of the intra-predicted block into partitions;

determine a partition-specific intra-prediction for each partition of the intra-predicted block using the intra-prediction mode;

decode, from the data stream, for each partition of the intra-predicted block, a partition-specific transform domain residual;

determine whether, for each partition of the intra-predicted block, no non-zero transform coefficient lies outside a respective area;

responsive to a determination that, for each partition of the intra-predicted block, no non-zero transform coefficient lies outside the respective area, decode a secondary transform index for the intra-predicted block from the data stream, wherein the secondary transform index indicates one of: 1) not using an inverse secondary transform, 2) using a first inverse secondary transform or 3) using a second inverse secondary transform;

determine a partition-specific spatial domain residual from the partition-specific transform domain residual for each partition of the intra-predicted block; and

reconstruct each partition of the intra-predicted block using the partition-specific spatial domain residual and the partition-specific intra-prediction.

2 . The video decoder of claim 1 , wherein the partition-specific spatial domain residual is computed based on the secondary transform index.

3 . The video decoder of claim 1 , configured to:

responsive to a determination that for at least one partition of the intra-predicted block there is a non-zero transform coefficient outside the respective area, not decode the secondary transform index from the data stream and infer that no secondary transform is used for the intra-predicted block.

4 . The video decoder of claim 1 , wherein:

the first inverse secondary transform or the second inverse secondary transform is further indicated by the intra-prediction mode.

5 . A method of video decoding, the method comprising:

decoding, from a data stream, an intra-prediction mode for an intra-predicted block of a picture;

decoding, from the data stream, information on a partitioning of the intra-predicted block into partitions;

determining a partition-specific intra-prediction for each partition of the intra-predicted block using the intra-prediction mode;

decoding, from the data stream, for each partition of the intra-predicted block, a partition-specific transform domain residual;

determining whether, for each partition of the intra-predicted block, no non-zero transform coefficient lies outside a respective area;

responsive to a determination that, for each partition of the intra-predicted block, no non-zero transform coefficient lies outside the respective area, decoding a secondary transform index for the intra-predicted block from the data stream, wherein the secondary transform index indicates one of: 1) not using an inverse secondary transform, 2) using a first inverse secondary transform or 3) using a second inverse secondary transform;

determining a partition-specific spatial domain residual from the partition-specific transform domain residual for each partition of the intra-predicted block; and

reconstructing each partition of the intra-predicted block using the partition-specific spatial domain residual and the partition-specific intra-prediction.

6 . The method of claim 5 , wherein the partition-specific spatial domain residual is determined based on the secondary transform index.

7 . The method of claim 5 , comprising:

responsive to a determination that for at least one partition of the intra-predicted block there is a non-zero transform coefficient outside the respective area, not decoding the secondary transform index from the data stream and inferring that no secondary transform is used for the intra-predicted block.

8 . The method of claim 5 , wherein:

the first inverse secondary transform or the second inverse secondary transform is further indicated by the intra-prediction mode.

9 . A video encoder, the video encoder configured to:

encode, into a data stream, an intra-prediction mode for an intra-predicted block of a picture;

encode, into the data stream, information on a partitioning of the intra-predicted block into partitions;

determine a partition-specific intra-prediction for each partition of the intra-predicted block using the intra-prediction mode;

encode, into the data stream, a partition-specific transform domain residual for each partition of the intra-predicted block;

determine whether, for each partition of the intra-predicted block, no non-zero transform coefficient lies outside a respective area; and

responsive to a determination that, for each partition of the intra-predicted block, no non-zero transform coefficient lies outside the respective area, encode a secondary transform index for the intra-predicted block into the data stream, wherein the secondary transform index indicates one of: 1) not using an inverse secondary transform, 2) using a first inverse secondary transform or 3) using a second inverse secondary transform.

10 . The video encoder of claim 9 , configured to:

responsive to a determination that for at least one partition of the intra-predicted block there is a non-zero transform coefficient outside the respective area, not encode the secondary transform index and infer that no secondary transform is used for the intra-predicted block.

11 . The video encoder of claim 9 , wherein:

the first inverse secondary transform or the second inverse secondary transform is further indicated by the intra-prediction mode.

12 . A method of video encoding, the method comprising:

encoding, into a data stream, an intra-prediction mode for an intra-predicted block of a picture;

encoding, into the data stream, information on a partitioning of the intra-predicted block into partitions;

determining a partition-specific intra-prediction for each partition of the intra-predicted block using the intra-prediction mode;

encoding, into the data stream, a partition-specific transform domain residual for each partition of the intra-predicted block;

determining whether, for each partition of the intra-predicted block, no non-zero transform coefficient lies outside a respective area; and

responsive to a determination that, for each partition of the intra-predicted block, no non-zero transform coefficient lies outside the respective area, encoding a secondary transform index for the intra-predicted block into the data stream, wherein the secondary transform index indicates one of: 1) not using an inverse secondary transform, 2) using a first inverse secondary transform or 3) using a second inverse secondary transform.

13 . The method of claim 12 , comprising:

responsive to a determination that for at least one partition of the intra-predicted block there is a non-zero transform coefficient outside the respective area, not encoding the secondary transform index and inferring that no secondary transform is used for the intra-predicted block.

14 . The method of claim 12 , wherein:

the first inverse secondary transform or the second inverse secondary transform is further indicated by the intra-prediction mode.

Assignments (1)

Priority Claims (3)

Continuity (3)

References Cited (18)