Point Cloud Occupancy Map Generation for Quality Preservation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for encoding and decoding point cloud data, particularly the video-based approach, can lead to deterioration in subjective quality due to the loss of unwanted points and incorrect representation of three-dimensional structures when using fixed filters for occupancy maps.
Innovation Solution
An information processing apparatus and method that generates an occupancyMap indicating the presence or absence of projection images of a point cloud on a two-dimensional plane based on positional relationships in three-dimensional space, allowing for the extraction and generation of accurate geometry and attribute data, and controlling filtering processes to prevent point loss by considering overlapping patch boundaries and neighboring points.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If filtering is performed on occupancyMap to reduce remainder, then encoding efficiency is improved, but points that should not be dropped are lost and subjective quality deteriorates
Solution Approach 1:
The patent performs preliminary classification of points into inside points and outside points before filtering. By pre-identifying which points are truly inside the object boundary using depth buffer comparison, the filtering process can safely remove only outside points without affecting inside points, thus maintaining quality while improving encoding efficiency.
Solution Approach 2:
The patent introduces depth buffer data as an intermediary to accurately determine point classification. By comparing point depths with depth buffer values at projected positions, the system creates an accurate classification mechanism that prevents misclassification, thereby avoiding quality deterioration while enabling effective filtering.
2Speed
If fixed filter is applied to occupancyMap, then processing speed is improved, but three-dimensional structure representation becomes incorrect
Solution Approach 1:
The patent applies different processing treatments to different regions of the occupancyMap based on point classification. Inside points retain their original values for accurate 3D structure representation, while outside points are filtered to zero. This localized differentiation maintains structural accuracy in critical regions while enabling speed improvements through selective filtering.
3Quantity of substance
If points are dropped to reduce data size, then compression ratio is improved, but fidelity of point cloud representation deteriorates
Solution Approach 1:
The patent uses depth buffer feedback to verify point classification accuracy. By continuously comparing point depths with depth buffer values during the filtering process, the system ensures that only points that should be removed are dropped, maintaining fidelity while achieving compression through selective point removal.
Data Source
AI summary
The present disclosure relates to an information processing apparatus and an information generation method that enable suppression of deterioration in subjective quality of a point cloud.There is generated an occupancyMap indicating, for each local area of a frame image, presence or absence of a projection image of a point cloud on a two-dimensional plane in accordance with a positional relationship of points in a three-dimensional space, while the point cloud represents an object having a three-dimensional shape as a set of the points. The present disclosure can be applied to, for example, an information processing apparatus, an encoding device, a decoding device, or the like.