Compressed Motion Vector Control for High-Resolution Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image/video compression technologies face challenges in efficiently compressing high-resolution, high-quality images/video with varying characteristics, leading to increased transmission and storage costs, particularly when motion vectors exceed 16 or 18 bits, necessitating improved methods to reduce the number of bits in motion vector representation.
Innovation Solution
A method and apparatus that derive a temporary reference motion vector, apply shifting operations to compress it, and generate a reference motion vector for the current block, reducing the total number of bits in its binary code representation, while applying shifting to affine control point motion vectors and controlling motion vector precision by removing Most Significant Bits (MSB) and Least Significant Bits (LSB).
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-resolution, high-quality image/video is transmitted or stored using existing technologies, then image quality is maintained, but transmission cost and storage cost increase due to increased amount of information
Solution Approach 1:
The patent extracts and removes redundant information from video data through various compression techniques including motion vector compression, transform coding, and quantization. By separating and eliminating unnecessary data components while preserving essential visual information, the system reduces the total amount of information that needs to be transmitted or stored.
Solution Approach 2:
The patent changes parameters of video data representation through quantization of transform coefficients, adjustment of motion vector precision, and variable bitrate encoding. These parameter changes allow the system to maintain acceptable image quality while significantly reducing the information quantity by representing data more efficiently.
2Measurement precision
If motion vectors are represented with high precision to maintain compression accuracy, then prediction accuracy is improved, but the number of bits required to represent motion vectors increases
Solution Approach 1:
The patent applies different precision levels to different parts of the motion data structure. Motion vectors are compressed by removing least significant bits (LSB) while preserving most significant bits (MSB), and affine control point motion vectors use differential encoding with selective bit removal. This local quality approach maintains prediction accuracy for large motions while reducing bit representation for minor variations.
Solution Approach 2:
The patent applies partial compression by selectively removing only certain bits (LSB) from motion vectors rather than compressing the entire vector. This partial action maintains sufficient prediction accuracy for most cases while reducing the bit count, allowing the system to achieve acceptable compression with less than full precision.
3Measurement precision
If affine control point motion vectors are used to improve prediction accuracy for complex motion, then prediction accuracy is improved, but the complexity of motion vector processing increases
Solution Approach 1:
The patent segments the motion vector processing into distinct components: control point motion vectors and differential motion vectors. By dividing the affine transformation into manageable parts that can be independently encoded and processed, the system reduces processing complexity while maintaining the predictive benefits of affine modeling.
Solution Approach 2:
The patent performs preliminary computation of affine control point motion vectors before encoding, storing pre-calculated values that can be directly used during decoding. This preliminary action eliminates the need for complex real-time affine transformations during the decoding process, significantly reducing processing complexity while maintaining prediction accuracy.
Data Source
AI summary
An image decoding method performed by the decoding apparatus according to this document comprising deriving a temporary reference motion vector for a current block based on a reference candidate block of the current block, deriving a reference motion vector by compressing the temporary reference motion vector based on a shifting operation, deriving a motion vector for the current block based on the reference motion vector, deriving prediction samples for the current block based on the motion vector for the current block, and generating reconstructed samples for the current block based on the prediction samples for the current block, wherein a total number of bits of the binary code representing the reference motion vector is less than a total number of bits of the binary code representing the temporary reference motion vector.