Adaptive HDR Pixel Filtering for Noise-Aware Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Noise in video content affects quality and encoding efficiency, necessitating improved methods to filter and compress high dynamic range (HDR) content.
Innovation Solution
Pixels are converted to a more perceptually uniform color space (e.g., CIE 1976 L*u*v*) and filtered based on surrounding intensity and complexity levels, using filter weights to adjust pixel colors and reduce noise before encoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If pixels are filtered to remove noise, then content quality is improved, but processing time and computational complexity increase
Solution Approach 1:
The patent applies different filtering strengths to different regions of the image based on their characteristics. Regions with higher noise levels receive stronger filtering, while regions with lower noise levels receive weaker filtering. This local adaptation allows quality improvement in noisy areas without unnecessarily processing entire images, thus reducing overall processing time.
Solution Approach 2:
The filtering process dynamically adjusts filter weights and parameters based on real-time analysis of image characteristics such as intensity levels, complexity, and noise patterns. This dynamic adaptation enables the system to optimize processing time by applying computational resources only where and when needed, rather than using fixed processing intensity throughout.
2Reliability
If filter weights are increased to remove more noise, then content quality is improved, but compression efficiency decreases
Solution Approach 1:
The patent changes filtering parameters such as filter weights, intensity thresholds, and complexity thresholds based on the specific characteristics of each image region and the desired compression level. By adjusting these parameters dynamically, the system achieves optimal balance between noise removal and compression efficiency, preventing over-filtering that would reduce compression efficiency.
Solution Approach 2:
The system applies partial filtering rather than maximum filtering throughout the entire image. By using selective filtering that targets only the most noisy regions with the highest filter weights, the system achieves sufficient quality improvement without the excessive processing that would harm compression efficiency.
3Measurement precision
If color space conversion is applied to improve filtering accuracy, then noise reduction is improved, but processing complexity increases
Solution Approach 1:
The patent segments the color space conversion process into selective regions rather than applying it uniformly to the entire image. By first analyzing image characteristics and then applying color space conversion only to regions where it is most beneficial for noise reduction, the system improves filtering accuracy where needed while reducing overall processing complexity.
Solution Approach 2:
The system performs preliminary analysis of image characteristics such as intensity levels and complexity before applying color space conversion and filtering. This preliminary action allows the system to determine in advance where color space conversion will be most effective for noise reduction, avoiding unnecessary computational complexity in regions where it would provide minimal benefit.
Data Source
AI summary
Systems, apparatuses, and methods are described for filtering and/or removing defects from content, such as high dynamic range (HDR) content. A plurality of parameters for filtering one or more pixels may be determined. The parameter(s) may be used to determine one or more filter weights, and the filter weight(s) may be applied to one or more pixels and one or more corresponding prior pixels to generate one or more filtered pixels. The filtered content and/or pixels thereof may later be encoded for storage and/or transmission to users.