US 10,382,767 B2 · Video coding using frame rotation

Abstract

Video frames are rotated for encoding and decoding blocks therein. During an encoding process, a number of degrees to rotate a video frame is determined. The video frame is rotated to produce rotated pixel values. A prediction residual is then generated for a block to be encoded based on the rotated pixel values. That block is encoded based on the prediction residual. During a decoding process, an encoded block is decoded to produce inverse transform pixel values, and a number of degrees to inverse rotate an encoded video frame is determined. The encoded video frame is inverse rotated to produce inverse rotated pixel values corresponding to the inverse transform pixel values. The inverse rotated pixel values are then outputted to an output video stream. One or more syntax elements indicating the number of degrees can be encoded to a bitstream during encoding and decoded from the bitstream during decoding.

Claims20

Dependent claims

A method for encoding a current block of a video frame to an encoded bitstream, the method comprising:

determining a number of degrees by which to rotate at least a portion of the video frame, wherein an object is located on a block boundary within the portion of the video frame;
rotating at least the portion of the video frame according to the determined number of degrees to produce rotated pixel values for at least the portion of the video frame, wherein at least some of the rotated pixel values correspond to the object, wherein rotating at least the portion of the video frame causes the object to no longer be located on the block boundary;
generating a prediction residual for the current block based on the rotated pixel values;
and
encoding the current block to the encoded bitstream based on the prediction residual.

2 more elements

A method for decoding an encoded video frame to an output video stream, the method comprising:

decoding the encoded video frame from an encoded bitstream to produce inverse transform pixel values;
determining a number of degrees by which to inverse rotate at least a portion of the decoded video frame, wherein an object within the portion of the decoded video frame is located on a block boundary;
inverse rotating at least the portion of the decoded video frame according to the determined number of degrees to produce inverse rotated pixel values for the inverse transform pixel values corresponding to at least the portion of the decoded video frame, wherein at least some of the inverse rotated pixel values correspond to the object, wherein inverse rotating at least the portion of the decoded video frame causes the object to be relocated on the block boundary;
and
outputting the inverse rotated pixel values to the output video stream.

2 more elements

An apparatus for decoding an encoded video frame to an output video stream, the apparatus comprising:

a processor configured to execute instructions stored in a non-transitory storage medium to:
decode the encoded video frame from an encoded bitstream to produce inverse transform pixel values;
determine a number of degrees by which to inverse rotate at least a portion of the decoded video frame, wherein an object within the portion of the decoded video frame is not located on a block boundary;
inverse rotate at least the portion of the decoded video frame according to the determined number of degrees to produce inverse rotated pixel values for the inverse transform pixel values that correspond to at least the portion of the decoded video frame, wherein at least some of the inverse rotated pixel values correspond to the object, wherein inverse rotating at least the portion of the decoded video frame causes the object to be relocated on the block boundary;
and
output the inverse rotated pixel values to the output video stream.

3 more elements

Dependencies are inferred from the claim text.

Video coding using frame rotation

Life of this patent

Video coding using frame rotation

Life of this patent

Abstract

Claims20