V-PCC-Guided Point Cloud Interpolation Through Multi-View 2D Projection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for interpolating point clouds face challenges in computation time and accuracy, particularly when increasing density, and 2D interpolation methods can generate incorrect points in 3D spaces.
Innovation Solution
A technology that applies 2D interpolation based on 3D spatial information using a V-PCC decoder, patch searcher, boundary point searcher, and boundary interpolation point generator to generate new points for occupied and unoccupied regions, utilizing 2D and 3D spatial information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If 3D operation is used to increase point cloud density, then the density can be increased, but computation time increases significantly due to complicated computation process
Solution Approach 1:
The patent projects 3D point cloud data onto 2D planes (front view, side view, top view) to perform interpolation operations in 2D space instead of 3D space. This dimensionality reduction simplifies the computation process while maintaining the ability to generate new points that are then mapped back to 3D coordinates, thereby reducing computation time while achieving density increase.
Solution Approach 2:
The patent divides the 3D point cloud into multiple 2D projection views (front, side, top) and processes each view separately. By segmenting the complex 3D interpolation problem into simpler 2D sub-problems, the overall computation complexity is reduced while still achieving comprehensive point cloud densification across all spatial dimensions.
2Device complexity
If 2D interpolation method is applied directly, then computation is simplified, but incorrect points are generated when points are spaced apart in 3D space but close in 2D space
Solution Approach 1:
The patent uses multiple 2D projection views (front, side, top) to represent the same 3D space from different angles. By performing interpolation in multiple 2D views and then combining the results in 3D space, the method maintains point location accuracy while keeping each individual interpolation process simple. The multi-view approach ensures that points spaced apart in 3D but close in one 2D view are properly distinguished through their positions in other views.
Solution Approach 2:
The patent introduces 2D projection views as intermediary representations between the original 3D point cloud and the final interpolated 3D points. These 2D projections serve as intermediate steps that simplify the interpolation computation while preserving spatial relationships, and the final 3D points are reconstructed by combining information from multiple intermediary views.
Data Source
AI summary
Disclosed are an apparatus for interpolating a point cloud and a method thereof. More particularly, a technology for interpolating a point cloud based on V-PCC decoding information and a 2D interpolation technology is disclosed.