Video frame drift correction
AssigneeHewlett-Packard Development Company, L.P.
InventorsMark Shaw, Jan P. Allebach, Edward J. Delp
Life of this patent
Estimating the term…
Abstract
In example implementations, a method executed by a processor is provided. The method determines an amount of video information that is lost in a video frame due to compression. A drift correction is applied to the video frame to add back a percentage of the amount of video information that is lost. The video frame is encoded with the drift correction.
Claims13
A method, comprising:
- determining, by a hardware device having a processor, a difference between a predicted macroblock of a video frame of a plurality of video frames between a first reference frame and a second reference frame and an actual macroblock of the video frame;
- determining, by the processor, an amount of information that is lost in the video frame due to compression based on the difference and a perceptibility of each pixel of the video frame;
- applying, by the processor, a drift correction to the video frame only when the video frame is one of a subset of video frames of the plurality of video frames beginning at time equal to (i+n)/2, where i is a time of the first reference frame and where n is a time of the second reference frame to add back a percentage of the amount of video information that is lost beginning at the time equal to (i+n)/2;
- and
- encoding, by the processor, the video frame with the drift correction to improve the image quality.
An apparatus, comprising:
- a processor,
- a machine-readable storage medium storing instructions executable by the processor to:
- increase a compression gain of a video;
- determine a difference between a predicted macroblock of a video frame of a plurality of video frames of the video and an actual macroblock of the video frame;
- determine an amount of information that is lost in the video frame due to compression based on the difference and a perceptibility of the video frame;
- apply a drift correction to the video frame only when the video frame is one of a subset of the plurality of video frames between a first I-frame and a second I-frame beginning at a time equal to (i+n)/2, where i is a time of the first I-frame and where n is a time of the second I-frame following the first I-frame, wherein to apply the drift correction includes to add back a percentage of the amount of information that is lost beginning at the time equal to (i+n)/2;
- perform a discrete cosine transform;
- and
- compress the plurality of video frames based on the compression gain and the drift correction to improve the image quality.
A non-transitory computer-readable storage medium encoded with instructions executable by a processor to improve a quality of an image, the computer-readable storage medium comprising:
- instructions to determine a residue of a video frame of a plurality of video frames between a first I-frame and a second I-frame from a difference between a predicted macroblock of the video frame and an actual macroblock of the video frame;
- instructions to subtract a perceptibility weighted delta of the video frame from the residue of the video frame to calculate a delta loss of the video frame;
- instructions to apply a drift correction to the delta loss of the video frame only when the video frame is one of a subset of video frames of the plurality of video frames, wherein the subset of video frames begins at a time equal to (i+n)/2, where i is a time of the first I-frame and where n is a time of the second I-frame;
- and
- instructions to generate a modified residue of the video frame based on the perceptibility weighted delta of the video frame plus the delta loss frame that includes the drift correction to improve an image quality.
Dependencies are inferred from the claim text.