Voxel Data Compression Using Distribution Pattern Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
General evaluation indices like PSNR do not accurately reflect subjective image quality, leading to noticeable deterioration in decoded point cloud data when used for merging nodes in voxel data compression.
Innovation Solution
An information processing apparatus and method that calculate a correlation degree based on subjective features of the distribution pattern of voxel data, comparing and merging voxel data blocks based on their distribution patterns to reduce redundancy and maintain subjective image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If general evaluation indices like PSNR are used for merging nodes in voxel data compression, then encoding efficiency is improved, but subjective image quality deteriorates
Solution Approach 1:
The patent changes the evaluation parameter from general objective metrics (PSNR) to subjective quality metrics that better reflect human perception. The correlation degree calculation unit computes similarity based on distribution patterns of voxel values, which correlates better with subjective image quality while maintaining encoding efficiency through optimized node merging decisions.
Solution Approach 2:
The patent replaces the traditional mechanical evaluation system (PSNR calculation) with a new evaluation mechanism based on distribution pattern correlation. This substitution uses statistical properties of voxel value distributions rather than simple pixel-wise error metrics, achieving better alignment with subjective quality assessment.
2Loss of information
If node merging is performed based on general correlation metrics, then data compression ratio is improved, but decoded point cloud quality deteriorates
Solution Approach 1:
The patent changes the correlation measurement parameter from general metrics to distribution pattern-based metrics. By evaluating the correlation of voxel value distributions rather than simple numerical differences, the system achieves better compression ratios while preserving decoded point cloud quality that aligns with subjective assessment.
Solution Approach 2:
The patent introduces feedback mechanisms where the correlation degree calculation results directly influence node merging decisions. The system continuously evaluates distribution pattern correlations and adjusts merging strategies accordingly, ensuring that compression operations maintain decoded quality while improving compression ratios.
Data Source
AI summary
The present disclosure relates to an information processing apparatus and a method that allows for suppression of a decrease in encoding efficiency. In comparison of voxel data resulting from quantization of point cloud data, a correlation degree of a distribution pattern between voxel data to be compared is calculated on the basis of a subjective feature of the distribution pattern of values of the voxel data. The present disclosure is applicable to an information processing apparatus, an image processing apparatus, an electronic device, an information processing method, a program, or the like, for example.