Adaptive False Contouring Prevention in Layered VDR Image Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Layered coding of images with extended dynamic range often results in false contouring artifacts due to quantization, which existing methods fail to completely eliminate without increasing computational complexity or degrading compression efficiency.
Innovation Solution
An adaptive quantization approach that uses a dynamic range compression function with adjustable parameters to minimize false contouring artifacts by iteratively adjusting the compression settings based on false contour detection metrics, ensuring the number of perceptually visible false contours is reduced.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If quantization is applied in layered VDR coding to compress the image data, then compression efficiency is improved, but false contouring artifacts are generated
Solution Approach 1:
The patent applies adaptive quantization where the quantization parameters are dynamically adjusted based on local image characteristics such as gradient magnitude and variance. By changing the quantization step size adaptively across different regions of the image, the method maintains compression efficiency while reducing false contouring artifacts in critical areas.
Solution Approach 2:
The patent implements region-specific quantization strategies where different quantization parameters are applied to different regions of the image based on their visual importance and complexity. High-gradient regions receive finer quantization to prevent false contours, while low-detail regions use coarser quantization to maintain compression efficiency.
2Productivity
If the number of quantization levels is reduced to increase compression, then compression efficiency is improved, but artifacts appear at block boundaries
Solution Approach 1:
The patent employs dynamic quantization parameter adjustment where the quantization levels are not fixed but adaptively determined based on local image statistics. This dynamic approach allows the system to use fewer quantization levels in compressible regions while maintaining higher precision at block boundaries where artifacts are more likely to occur.
Solution Approach 2:
The patent performs preliminary analysis of image blocks to identify regions prone to block boundary artifacts before applying quantization. By pre-processing the image to detect edge regions and high-variation areas, the system can pre-assign appropriate quantization parameters that prevent artifact formation before compression is applied.
3Object-generated harmful factors
If adaptive quantization is applied to prevent false contours, then artifact reduction is improved, but computational complexity increases
Solution Approach 1:
The patent applies adaptive quantization selectively only to regions of the image where false contouring is likely to occur, such as regions with high gradient magnitude or low variance. By limiting the adaptive processing to only these critical regions rather than the entire image, the computational complexity is significantly reduced while still achieving effective artifact prevention.
4Object-generated harmful factors
If iterative adjustment of compression parameters is performed to minimize false contours, then artifact prevention is improved, but processing time increases
Solution Approach 1:
The patent uses simplified, computationally inexpensive metrics for evaluating false contouring potential and for adjusting quantization parameters. Rather than employing complex iterative optimization algorithms, the method uses direct calculations based on local image statistics that can be computed quickly and discarded after use, achieving effective artifact prevention with minimal processing overhead.
Data Source
Figure 1
Figure 2~3
Figure 4~6A
AI summary
An encoder receives a sequence of images in extended or visual dynamic range (VDR). For each image, a dynamic range compression function and associated parameters are selected to convert the input image into a second image with a lower dynamic range. Using the input image and the second image, a residual image is computed. The input VDR image sequence is coded using a layered codec that uses the second image as a base layer and a residual image that is derived from the input and second images as one or more residual layers. Using the residual image, a false contour detection method (FCD) estimates the number of potential perceptually visible false contours in the decoded VDR image and iteratively adjusts the dynamic range compression parameters to prevent or reduce the number of false contours. Examples that use a uniform dynamic range compression function are also described.