Anchor Point Cloud Compression Color Leaking Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Anchor Point Cloud Compression experiences significant color distortion, known as color leaking, even at low compression ratios due to chroma subsampling in the JPEG process, which affects the accuracy of 3D point cloud data.
Innovation Solution
Implementing 1D subsampling of chroma components instead of 2D decimation and interpolation, and using JPEG-4:4:4 sampling instead of JPEG-4:2:0 to reduce color leaking, with an adaptive approach selecting between these methods based on performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If chroma subsampling is applied in JPEG compression to reduce data size, then compression ratio is improved, but color distortion (color leaking) increases
Solution Approach 1:
The patent transforms the 2D chroma subsampling operation into a 1D operation by reorganizing the chroma component data from a two-dimensional array into a one-dimensional time series. This dimensional change allows for different decimation strategies that preserve temporal correlations while reducing spatial artifacts, thereby mitigating color leaking while maintaining compression efficiency.
Solution Approach 2:
The patent introduces adaptive parameters including the decimation factor (controlling the degree of subsampling) and the choice of interpolation method (linear, cubic, or spline). These parameters can be adjusted based on the specific characteristics of the point cloud data and the desired balance between compression ratio and color accuracy, allowing optimization of the trade-off between the two conflicting objectives.
2Productivity
If 2D decimation and interpolation is used for chroma subsampling, then compression efficiency is improved, but color leaking artifacts increase
Solution Approach 1:
The patent converts the 2D chroma subsampling problem into a 1D problem by linearizing the chroma component array. This transformation changes the neighborhood relationships used for interpolation, reducing the propagation of color artifacts while maintaining compression efficiency. The 1D approach processes chroma values along a single dimension rather than applying 2D filtering that causes color leaking.
Solution Approach 2:
The patent replaces the traditional 2D spatial filtering mechanism with a 1D temporal sequence processing approach. By treating the chroma component as a time series rather than a spatial image, the patent substitutes the mechanical 2D decimation-interpolation system with a 1D signal processing system that is less prone to generating color leaking artifacts.
3Quantity of substance
If JPEG-4:2:0 sampling is used, then data reduction is improved, but visual quality of decoded point clouds deteriorates
Solution Approach 1:
The patent implements a dynamic approach where the chroma decimation factor can be adjusted based on the complexity of the scene and the desired quality level. Rather than fixed 4:2:0 subsampling, the system can adaptively choose between different decimation factors (e.g., 2:1 or 4:1) and interpolation methods to optimize the balance between data reduction and visual quality for different point cloud contents.
Solution Approach 2:
The patent changes the sampling parameters from the fixed JPEG-4:2:0 standard to an adaptive scheme where the chroma subsampling factor and reconstruction method can be modified. This allows the system to maintain higher visual quality by using less aggressive subsampling when needed, while still achieving significant data reduction when appropriate, thus improving overall visual quality compared to fixed 4:2:0 sampling.
Data Source
AI summary
The color coding approach adopted by the Anchor Point Cloud Compression (PCC) leads to a significant color distortion, even at very low compression ratios. This color distortion is also referred to as color leaking. Two approaches are able to mitigate the color leaking. Both approaches reduce the color leaking significantly, while one of them (e.g., 1D subsampling) is more robust and preferred in most cases. A more general approach is an adaptive selection between these two approaches.