Video encoding method and apparatus, computer device, and storage medium

Patent metadata

Field	Value
Publication number	US-12355941-B2
Application number	US-202217678931-A
Country	US
Kind code	B2
Filing date	Feb 23, 2022
Priority date	Mar 2, 2020
Publication date	Jul 8, 2025
Grant date	Jul 8, 2025

Abstract

A method of video encoding is described. The method includes segmenting original video data to obtain an original video segment including multiple video images. Video content analysis is performed on the original video segment to obtain a video image processing parameter corresponding to the original video segment. Image processing is performed on a video image in the multiple video images in the original video segment based on the video image processing parameter to obtain a processed video segment. An encoding parameter of the processed video segment can be determined based on image feature data of the processed video segment. The processed video segment can be encoded based on the encoding parameter to obtain an encoded video segment.

First claim

What is claimed is:

1. A method of video encoding, comprising:
segmenting, by processing circuitry, original video data to obtain an original video segment including multiple video images;
inputting the original video segment and target image quality data to a video analysis model;
obtaining a video image processing parameter corresponding to the original video segment from a plurality of video image processing parameters and outputted by the video analysis model;
performing, by the processing circuitry, image processing on a video image in the multiple video images in the original video segment based on the video image processing parameter to obtain a processed video segment;
determining, by the processing circuitry, an encoding parameter of the processed video segment from a plurality of encoding parameters based on image feature data of the processed video segment, the image feature data of the processed video segment being extracted from the processed video segment; and
encoding, by the processing circuitry, the processed video segment based on the encoding parameter and respective associations between the plurality of video image processing parameters and the plurality of encoding parameters to obtain an encoded video segment.

2. The method according to claim 1 , wherein
the video analysis model is a machine learning model obtained through training performed by using a video segment sample, target image quality data of the video segment sample, and a video image processing parameter of the video segment sample.

3. The method according to claim 2 , wherein the target image quality data comprises an image quality level indicating a target image quality for encoding the original video segment.

4. The method according to claim 3 , wherein the video image processing parameter comprises at least one of the following:
a target frame rate, a target quantization bit depth, and a luminance adjustment curve.

5. The method according to claim 4 , wherein the performing the image processing comprises:
in response to that the video image processing parameter comprises the target frame rate and the target frame rate is higher than a frame rate of the original video segment, adjusting the frame rate of the original video segment through frame rate upsampling to obtain the processed video segment;
in response to that the video image processing parameter comprises the target frame rate and the target frame rate is lower than the frame rate of the original video segment, adjusting the frame rate of the original video segment through frame rate downsampling to obtain the processed video segment; and
in response to that the video image processing parameter comprises the target frame rate and the target frame rate is equal to the frame rate of the original video segment, maintaining the frame rate of the original video segment to obtain the processed video segment.

6. The method according to claim 4 , wherein the performing the image processing comprises:
in response to that the video image processing parameter comprises the target quantization bit depth and the target quantization bit depth is lower than a quantization bit depth of the original video segment, performing downsampling quantization on the original video segment to reduce the quantization bit depth of the original video segment to obtain the processed video segment;
in response to that the video image processing parameter comprises the target quantization bit depth and the target quantization bit depth is higher than the quantization bit depth of the original video segment, performing reverse high-precision inverse quantization on the original video segment to increase the quantization bit depth of the original video segment to obtain the processed video segment; and
in response to that the video image processing parameter comprises the target quantization bit depth and the target quantization bit depth is equal to the quantization bit depth of the original video segment, maintaining the quantization bit depth of the original video segment to obtain the processed video segment.

7. The method according to claim 4 , wherein the performing the image processing comprises:
in response to that the video image processing parameter comprises the luminance adjustment curve and a luminance range corresponding to the luminance adjustment curve is inconsistent with a luminance range of the original video segment, performing tone mapping on the original video segment based on the luminance adjustment curve to obtain the processed video segment; and
in response to that the video image processing parameter comprises the luminance adjustment curve and the luminance range corresponding to the luminance adjustment curve is consistent with the luminance range of the original video segment, maintaining a tone of the original video segment to obtain the processed video segment.

8. The method according to claim 1 , wherein the determining comprises:
inputting the image feature data of the processed video segment to an encoding parameter determining model to obtain the encoding parameter outputted by the encoding parameter determining model, wherein
the encoding parameter determining model is a machine learning model obtained through training performed by using an image feature data sample and an encoding parameter corresponding to the image feature data sample.

9. The method according to claim 8 , wherein the image feature data comprises at least one of the following:
a frame rate, a quantization bit depth, a maximum luminance, a minimum luminance, an image type, a motion vector, and target image quality data.

10. The method according to claim 9 , wherein the encoding parameter comprises a bit rate.

11. The method according to claim 1 , wherein the segmenting comprises:
segmenting the original video data based on a specified dimension of image property to obtain the original video segment, wherein
the specified dimension of image property comprises at least one of distribution characteristics of a dark part and a highlight part in an image, a track and a motion degree of a motion area, color distribution and intensity, and details of a picture.

12. The method according to claim 1 , wherein the method further comprises:
combining all encoded video segments in a corresponding segmentation sequence of the original video data to obtain encoded video data.

13. An apparatus of video encoding, comprising:
processing circuitry configured to
segment original video data to obtain an original video segment including multiple video images;
input the original video segment and target image quality data to a video analysis model;
obtain a video image processing parameter corresponding to the original video segment from a plurality of video image processing parameters and outputted by the video analysis model;
perform image processing on a video image in the multiple video images in the original video segment based on the video image processing parameter to obtain a processed video segment;
determine an encoding parameter of the processed video segment from a plurality of encoding parameters based on image feature data of the processed video segment, the image feature data of the processed video segment being extracted from the processed video segment; and
encode the processed video segment based on the encoding parameter and respective associations between the plurality of video image processing parameters and the plurality of encoding parameters to obtain an encoded video segment.

14. The apparatus according to claim 13 , wherein
the video analysis model is a machine learning model obta

Assignees

• Tencent Tech Shenzhen Co Ltd

Inventors

• Li Dalong

Classifications

• H04N19/186

  the unit being a colour or a chrominance component · CPC title

• H04N19/132

  Sampling, masking or truncation of coding units, e.g. adaptive resampling, frame skipping, frame interpolation or high-frequency transform coefficient masking · CPC title

• H04N19/124 Primary

  Quantisation · CPC title

• H04N19/117

  Filters, e.g. for pre-processing or post-processing (sub-band filter banks H04N19/635) · CPC title

• H04N19/85

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