Chroma Subblock Affine Prediction Across Multiple Chroma Formats
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies, such as the Versatile Video Coding (VTM) coder, struggle with supporting chroma formats other than 4:2:0, leading to crashes when encountering formats like 4:4:4, necessitating a solution to enhance coding performance for a variety of applications.
Innovation Solution
A method and apparatus for affine based inter prediction of chroma subblocks that derive motion vectors by determining chroma scaling factors and using luma motion vectors to improve accuracy, allowing support for multiple chroma formats, including 4:4:4, by averaging motion vectors of luma subblocks based on chroma format information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the VTM coder is designed to support only chroma format 4:2:0, then the coding process is simple and stable, but it crashes when encountering other chroma formats like 4:4:4
Solution Approach 1:
The motion vector derivation mechanism is designed to handle multiple chroma formats (4:2:0, 4:2:2, 4:4:4) through a unified approach. The system determines chroma scaling factors based on chroma format information and uses luma motion vectors to derive chroma motion vectors regardless of the specific chroma format, making the coder universal across different formats without requiring separate processing paths for each format.
Solution Approach 2:
The system adapts to different chroma formats by dynamically determining chroma scaling factors (SubWidthC and SubHeightC) based on the chroma format information. These scaling factors change according to the chroma format (e.g., 4:2:0 has different scaling than 4:4:4), allowing the motion vector derivation to adjust its parameters to match the specific chroma format being processed, thereby maintaining stability across diverse formats.
2Measurement precision
If chroma motion vectors are derived directly without using luma motion vectors, then the processing is faster, but the prediction accuracy decreases
Solution Approach 1:
The system performs preliminary action by deriving chroma motion vectors from luma motion vectors before actual chroma prediction. Luma motion vectors are obtained first (as they have higher accuracy due to larger block sizes and better signal-to-noise ratio), and then chroma motion vectors are derived from them through scaling and offset operations. This preliminary derivation using high-accuracy luma data improves the overall motion compensation accuracy for chroma.
Solution Approach 2:
Luma motion vectors serve as an intermediary to derive chroma motion vectors. Instead of directly estimating chroma motion vectors (which would be noisy and inaccurate), the system uses luma motion vectors as an intermediate step. The luma motion vectors are scaled and adjusted to obtain chroma motion vectors, leveraging the higher quality luma data to improve chroma prediction accuracy.
3Measurement precision
If chroma subblocks are processed independently without affine prediction, then the computational complexity is lower, but the prediction error increases
Solution Approach 1:
The chroma block is segmented into multiple chroma subblocks, and affine prediction is applied independently to each subblock. Each chroma subblock has its own motion vector derived from corresponding luma subblocks. This segmentation allows the system to capture local motion variations within the chroma block more accurately than a single uniform prediction would provide, improving prediction accuracy for complex motion patterns.
Solution Approach 2:
Different chroma subblocks are processed with locally adapted motion vectors derived from their corresponding luma subblocks. Each chroma subblock receives a motion vector that is specific to its local region, allowing the prediction to adapt to local motion characteristics. This local quality approach ensures that regions with different motion patterns are predicted accurately without being constrained by a single global motion model.
Data Source
AI summary
The present invention relates to method and apparatus for motion vector derivation for affine based inter prediction of chroma subblocks based on chroma format. The method includes: determining chroma scaling factors in horizontal and vertical directions based on chroma format information, wherein the chroma format information indicates a chroma format of a current picture which the current image block belongs to; determining a set(S) of luma subblocks of the luma block based on values of the chroma scaling factors; determining a motion vector for a chroma subblock of the chroma block based on motion vectors of one or more luma subblocks in the set(S) of luma subblocks.