Pixel level adaptive intra-smoothing

Patent metadata

Field	Value
Publication number	US-9930366-B2
Application number	US-201113267762-A
Country	US
Kind code	B2
Filing date	Oct 6, 2011
Priority date	Jan 28, 2011
Publication date	Mar 27, 2018
Grant date	Mar 27, 2018

Abstract

This disclosure describes intra-smoothing techniques for intra-coding of video data. In one example, a video encoder is configured to determine a plurality of prediction samples for intra-predictive coding of a block to be coded in a video frame, calculate a local statistic for a first prediction sample of the prediction samples, select a filter from a plurality of filters based on the local statistic, and apply the selected filter to the first prediction sample. The video encoder may also be configured to calculate at least one local statistic for each prediction sample of the prediction samples, select a filter from a plurality of different filters based on the at least one local statistic for each prediction sample, and apply each selected filter to the corresponding prediction sample to generate the filtered version of the prediction sample. Embodiments further include video decoders and method of decoding encoded video signals.

First claim

The invention claimed is:

1. A method of processing video data, the method comprising:
encoding a first block of video data;
calculating a local statistic for each pixel of a plurality of pixels in a second block of video data to be encoded, wherein the second block of video data forms at least a portion of a video frame, and wherein the local statistic comprises a variance calculated using values of two or more other pixels, the other pixels being included within the second block of video data and being pixels other than the pixel for which the local statistic is calculated;
obtaining prediction samples from the first block of video data;
selecting, for each respective pixel of the plurality of pixels in the second block of video data, between a respective filtered version of a prediction sample of the prediction samples from the first block of video data and a respective unfiltered version of the prediction sample based on the local statistic;
applying, for each respective pixel of the plurality of pixels, the respective selected version of the prediction sample to intra-predict the respective pixel; and
encoding the second block of video data using the intra-predicted pixels.

2. The method of claim 1 , further comprising:
calculating at least one local statistic for each prediction sample of the prediction samples;
for each prediction sample of the prediction samples, selecting a filter from a plurality of different filters based on the at least one local statistic for each prediction sample; and
applying each selected filter to the corresponding prediction sample to generate the filtered version of the prediction sample.

3. The method of claim 1 , wherein calculating the local statistic comprises calculating the local statistic based at least in part on one or more intensity values associated with one or more of the other pixels, or one or more positions associated with one or more of the other pixels.

4. The method of claim 1 , wherein calculating the local statistic for each respective pixel of the plurality of pixels in the second block of video data to be encoded further comprises calculating a local statistic for each respective pixel based on at least one local causal neighbor of the respective pixel.

5. The method of claim 4 , wherein the at least one local causal neighbor of the respective pixel comprises four pixels adjacent to the respective pixel.

6. The method of claim 1 , further comprising:
determining a threshold value for the local statistic, and
wherein selecting between a filtered version of a prediction sample of the prediction samples and an unfiltered version of the prediction sample further comprises:
selecting the unfiltered version of the prediction sample when a value of the local statistic is below the threshold value; and
selecting the filtered version of the prediction sample when the value of the local statistic is equal to or greater than the threshold value.

7. A device for encoding data, comprising:
a memory storing video data; and
one or more processors configured to:
encode a first block of video data;
calculate a local statistic for each pixel in a second block of video data to be encoded of a plurality of pixels, wherein the second block of video data forms at least a portion of a video frame, and wherein the local statistic comprises a variance calculated using values of two or more other pixels, the other pixels being included within the second block of video data and being pixels other than the pixel for which the local statistic is calculated;
obtain prediction samples from the first block of video data;
select, for each respective pixel of the plurality of pixels in the second block of video data, between a filtered version of a prediction sample of the prediction samples from the first block of video data and a respective unfiltered version of the prediction sample based on the local statistic;
apply, for each respective pixel of the plurality of pixels, the respective selected version of the prediction sample to intra-predict the respective pixel; and
encode the second block of video data to be encoded using the intra-predicted pixels.

8. The device of claim 7 , wherein the one or more processors are further configured to calculate at least one local statistic for each prediction sample of the prediction samples, select a filter from a plurality of different filters based on the at least one local statistic for each prediction sample of the prediction samples, and apply each selected filter to the corresponding prediction sample to generate the filtered version of the prediction sample.

9. The device of claim 8 , wherein, to calculate the local statistic, the one or more processors are configured to calculate the local statistic based at least in part one or more intensity values associated with one or more of the other pixels, or one or more positions associated with one or more of the other pixels.

10. The device of claim 7 , wherein, to calculate the local statistic for each respective pixel of the plurality of pixels in the second block of video data to be encoded, the one or more processors are configured to calculate a local statistic for each respective pixel based on at least one local causal neighbor of the respective pixel.

11. The device of claim 10 , wherein the at least one local causal neighbor of the respective pixel comprises four pixels adjacent to the respective pixel.

12. The device of claim 7 , wherein the one or more processors are further configured to determine a threshold value for the local statistic, and wherein, to select between a filtered version of a prediction sample of the prediction samples and an unfiltered version of the prediction sample, the one or more processors are configured to select the unfiltered version of the prediction sample when a value of the local statistic is below the threshold value and select the filtered version of the prediction sample when the value of the local statistic is equal to or greater than the threshold value.

13. A non-transitory computer-readable storage medium having stored thereon instructions that, when executed, cause a processor of a video encoding device to:
encode a first block of video data;
calculate a local statistic for each pixel of a plurality of pixels in a second block of video data to be encoded, wherein the second block of video data forms at least a portion of a video frame, and wherein the local statistic comprises a variance calculated using values of two or more other pixels, the other pixels being included within the second block of video data to be encoded and being pixels other than the pixel for which the local statistic is calculated;
obtain prediction samples from the first block of video data;
select, for each respective pixel of the plurality of pixels in the second block of video data, between a respective filtered version of a prediction sample of the prediction samples from the first block of video data and a respective unfiltered version of the prediction sample based on the local statistic;
apply, for each respective pixel of the plurality of pixels, the respective selected version of the prediction sample to intra-predict the respective pixel; and
encode the second block of video data to be encoded using the intra-predicted pixels.

14. The non-transitory computer-readable storage medium of claim 13 , wherein the instructions further cause the processor of the video encoding device to:
calculate at least one local statistic for each prediction sample of the prediction samples;
for each prediction sample of the prediction samples, select a filter from a plurali

Assignees

• Zheng Yunfei
• Coban Muhammed Zeyd
• Karczewicz Marta
• Qualcomm Inc

Inventors

• Zheng Yunfei
• Coban Muhammed Zeyd
• Karczewicz Marta

Classifications

• H04N19/00 Primary

  Methods or arrangements for coding, decoding, compressing or decompressing digital video signals · CPC title

• H04N19/105

  Selection of the reference unit for prediction within a chosen coding or prediction mode, e.g. adaptive choice of position and number of pixels used for prediction · CPC title

• H04N19/85 Primary

  using pre-processing or post-processing specially adapted for video compression · CPC title

• H04N19/182

  the unit being a pixel · CPC title

• H04N19/136

  Incoming video signal characteristics or properties · CPC title