VVC Block Division for ZeroOut Quality and Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In real-time encoding of images using versatile video coding (VVC), the selection of transformation sizes that apply ZeroOut can lead to a reduction in subjective image quality without considering the coding efficiency, as the high-frequency components are completely lost, impacting the decoded image's quality.
Innovation Solution
An information processing device and method that estimates the influence of ZeroOut on subjective image quality by analyzing quantization coefficients and controls block division based on this influence to set the block size, considering both coding efficiency and subjective image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If transformation size is set to apply ZeroOut to reduce circuit scale of decoder, then coding efficiency is improved, but high-frequency components are lost and subjective image quality is reduced
Solution Approach 1:
The patent applies different transformation sizes to different blocks within an image. Blocks with strong spatial correlation use larger transformation sizes with ZeroOut for efficient coding, while blocks with weak spatial correlation use smaller transformation sizes without ZeroOut to preserve high-frequency components. This local differentiation resolves the contradiction by allowing both coding efficiency and image quality optimization in different regions.
Solution Approach 2:
The patent dynamically determines the transformation size for each block based on the calculated correlation between adjacent pixels. The correlation threshold is used to adaptively select whether to apply ZeroOut or not, making the transformation size selection dynamic rather than fixed. This allows the system to optimize both coding efficiency and image quality based on local image characteristics.
2Productivity
If transformation size is set to maximize coding efficiency without considering image quality, then coding speed is improved, but subjective image quality is greatly reduced
Solution Approach 1:
The patent calculates the correlation between adjacent pixels before determining the transformation size. This preliminary correlation calculation enables the system to predict which blocks will benefit from ZeroOut and which will lose quality, allowing for pre-planned transformation size assignment that balances coding speed and image quality without requiring post-processing adjustments.
Solution Approach 2:
The patent changes the transformation size parameter based on the calculated pixel correlation values. By using the correlation threshold to switch between different transformation size options, the system dynamically adjusts the coding parameters to optimize both speed and quality. This parameter adaptation allows real-time encoding while maintaining acceptable image quality.
3Manufacturing precision
If correlation calculation is performed for each block to determine transformation size, then subjective image quality is improved, but processing time is increased
Solution Approach 1:
The patent applies correlation calculation and transformation size optimization only to blocks that have weak spatial correlation (where it matters most for image quality). Blocks with strong spatial correlation use default larger transformation sizes without individual correlation analysis. This partial application of the correlation analysis approach maintains image quality where needed while minimizing overall processing time.
Data Source
AI summary
The present disclosure relates to an information processing device and a method enabling suppression of reduction in subjective image quality while suppressing reduction in coding efficiency. An influence on a subjective image quality of a decoded image by applying ZeroOut to quantization coefficients corresponding to an image when coding the image is estimated, and division of a block is controlled on the basis of the estimated influence when setting a size of the block serving as a unit of processing in the coding of the image. The present disclosure may be applied to, for example, an information processing device, a coding device, an electronic device, an information processing method, a program or the like.