Geometric transformations for video compression
Decoding a current block of a current frame includes decoding, from a compressed bitstream, one or more syntax elements indicating that a geometric transformation is to be applied; applying the geometric transformation to at least a portion of the current frame to obtain a transformed portion; and obtaining a prediction of the current block based on the transformed portion and an intra-prediction mode.
1 . A method for decoding a current block of a current frame comprising:
decoding, from a compressed bitstream, one or more syntax elements indicating that a geometric transformation is to be applied to at least a portion of the current frame for intra-prediction purposes, wherein the geometric transformation operates in a pixel domain and produces a result in the pixel domain;
applying the geometric transformation to at least a portion of the current frame to obtain a transformed portion within the current frame; and
obtaining a prediction of the current block based on the transformed portion and an intra-prediction mode using reference pixels from the transformed portion of the current frame.
2 . The method of claim 1 , wherein the geometric transformation is a horizontal reflection and the at least the portion of the current frame comprises a neighboring block that is above the current block.
3 . The method of claim 1 , wherein the geometric transformation is a vertical reflection and the at least the portion of the current frame comprises a neighboring block that is left of the current block.
4 . The method of claim 1 , wherein the geometric transformation comprises a line-straightening transformation.
5 . The method of claim 4 , wherein the line-straightening transformation is based on a distortion model.
6 . The method of claim 5 , wherein an indicator of the distortion model is decoded from the compressed bitstream.
7 . The method of claim 1 , wherein the geometric transformation comprises a rotation operation.
8 . The method of claim 1 , wherein the at least the portion of the current frame comprises a tile of the current frame.
9 . A method for encoding a current block of a current frame comprising:
determining that the current block is to be encoded based on a geometric transformation to be applied to at least a portion of the current frame for intra-prediction purposes, wherein the geometric transformation operates in a pixel domain and produces a result in the pixel domain;
encoding, in a compressed bitstream, one or more syntax elements indicating the geometric transformation;
applying the geometric transformation to at least a portion of the current frame that includes the current block to obtain a transformed portion within the current frame; and
obtaining a prediction block for the current block based on the transformed portion and an intra-prediction mode using reference pixels from the transformed portion of the current frame.
10 . The method of claim 9 , wherein the geometric transformation is a horizontal reflection and the at least the portion of the current frame comprises a neighboring block that is above the current block.
11 . The method of claim 9 , wherein the geometric transformation is a vertical reflection and the at least the portion of the current frame comprises a neighboring block that is left of the current block.
12 . The method of claim 9 , wherein the geometric transformation comprises a line-straightening transformation.
13 . The method of claim 12 , wherein the line-straightening transformation is based on a distortion model.
14 . The method of claim 9 , wherein the geometric transformation comprises a rotation operation.
15 . The method of claim 9 , wherein the at least the portion of the current frame comprises a tile of the current frame.
16 . A device, comprising:
a processor, the processor configured to execute instructions to code a current block of a current frame, the instructions comprise to:
code one or more syntax elements indicating that a geometric transformation is to be applied to at least a portion of the current frame for intra-prediction purposes, wherein the geometric transformation operates in a pixel domain and produces a result in the pixel domain;
apply the geometric transformation to at least a portion of the current frame to obtain a transformed portion within the current frame; and
obtain a prediction of the current block based on the transformed portion and an intra-prediction mode using reference pixels from the transformed portion of the current frame.
17 . The device of claim 16 , wherein the geometric transformation comprises at least one of a horizontal, a vertical reflection, or a line-straightening transformation.
18 . The device of claim 16 , wherein the geometric transformation comprises a rotation operation.
19 . The device of claim 16 , wherein the geometric transformation is a horizontal reflection and the at least the portion of the current frame comprises a neighboring block that is above the current block.
20 . The device of claim 16 , wherein the geometric transformation is a vertical reflection and the at least the portion of the current frame comprises a neighboring block that is left of the current block.