3D Surface Extraction via Hierarchical Tree Structure for Memory Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current surface extraction methods in 3D reconstruction for AR applications require large amounts of depth data, leading to high memory usage and inefficiencies in processing.
Innovation Solution
A method that groups neighboring 3D child nodes into parent nodes, propagates distances, and recursively generates a tree structure to perform surface extraction, using the dual marching cube algorithm to efficiently extract surfaces while reducing memory usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional surface extraction methods are used with complete depth data, then surface extraction accuracy is improved, but memory usage increases significantly
Solution Approach 1:
The patent segments the depth data processing into two distinct phases: an encoding phase that creates a compact tree structure representation of the scene, and a decoding phase that reconstructs surface information only where needed. This segmentation allows accurate surface extraction while using minimal memory to store only the essential tree structure rather than complete depth maps for all voxels.
Solution Approach 2:
The patent performs preliminary encoding of the entire scene into a compact tree structure before actual surface extraction occurs. During this preliminary action, depth information is processed and stored in an optimized format that preserves all necessary geometric relationships while eliminating redundant data, enabling accurate subsequent surface extraction with minimal memory requirements.
2Reliability
If complete depth data is stored for all 3D child nodes, then surface extraction completeness is improved, but processing time increases
Solution Approach 1:
The patent extracts only the essential geometric information needed for surface extraction and stores it in a compact tree structure. Instead of retaining complete depth data for all possible 3D child nodes, the method extracts and stores only the boundary-surface-defining information in the tree, eliminating unnecessary data while maintaining extraction completeness.
Solution Approach 2:
The patent performs preliminary encoding that pre-processes the entire scene into an optimized tree structure before surface extraction. This preliminary action organizes all necessary geometric relationships in advance, enabling rapid query and extraction operations without requiring time-consuming processing during the actual surface extraction phase.
3Measurement precision
If depth data is propagated to all parent nodes in the tree structure, then surface extraction accuracy is improved, but memory usage increases
Solution Approach 1:
The patent applies local quality by storing depth propagation information only at specific nodes in the tree structure where it is actually needed for surface extraction, rather than uniformly at all nodes. The tree structure selectively preserves depth relationships locally at boundary-surface-defining nodes while omitting redundant information at internal nodes, maintaining accuracy where required while minimizing memory usage.
Data Source
AI summary
A surface extraction method that includes the steps outlined below is provided. (A) Raw input depth data of 3D child nodes are received that includes positions and distances of first 3D child nodes and positions of second 3D child nodes. (B) The neighboring 3D child nodes are grouped into 3D parent nodes. (C) For each 3D parent nodes, propagated distances of the second 3D child nodes are generated. (D) The 3D parent nodes are treated as the 3D child nodes to perform the steps (B)-(D) to generate a tree including a plurality levels of nodes such that a surface extraction processing is performed on the tree to extract at least one surface of the scene.