3D Scene Point Generation via Depth Difference Interpolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for processing 3D scenes using point clouds often result in erroneous reconstruction or missing points, leading to holes in the scene, especially when the viewpoint changes, due to inefficiencies in representing complex geometries like hair or fur.
Innovation Solution
A method and device that generate additional points in a 3D scene by comparing depth information of current pixels with adjacent pixels, creating additional points based on depth differences within specific thresholds, and interpolating texture and depth attributes to fill holes and increase point density.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If depth compression is applied to reduce data size, then storage efficiency is improved, but reconstruction accuracy deteriorates leading to holes in the 3D scene
Solution Approach 1:
The patent performs preliminary depth interpolation during the encoding phase by comparing depth values of current pixels with adjacent pixels and generating intermediate depth values. This preliminary action ensures that when the compressed data is decoded, the intermediate depth information is already available to prevent reconstruction holes, thus maintaining accuracy despite compression.
Solution Approach 2:
The patent introduces intermediate depth values as mediators between compressed depth data and the final 3D reconstruction. These intermediate values are generated by comparing current pixel depth with adjacent pixels and filling in missing depth information, acting as a bridge that preserves reconstruction accuracy while allowing data compression.
2Manufacturing precision
If point cloud density is increased to improve scene completeness, then representation accuracy is improved, but data processing complexity increases
Solution Approach 1:
The patent applies local quality by performing depth interpolation only in specific regions where depth discontinuities are detected between adjacent pixels. Instead of uniformly increasing point cloud density everywhere, the method selectively generates additional points only where needed to fill holes, thus improving scene completeness while minimizing unnecessary data processing complexity.
3Device complexity
If complex geometries like hair or fur are represented with traditional mesh formats, then surface representation is simplified, but geometric accuracy deteriorates
Solution Approach 1:
The patent segments the representation of complex geometries by using point cloud data instead of continuous mesh surfaces. Each point in the point cloud independently represents a specific location, allowing accurate representation of discontinuous structures like hair and fur that cannot be efficiently represented by traditional mesh elements, thus maintaining geometric accuracy while keeping the representation manageable.
Data Source
AI summary
A method and device for generating points of a 3D scene from a depth image. To reach that aim, depth information associated with a current pixel is compared with depth information associated with pixels spatially adjacent to the current pixel in the depth image. When the difference of depth between the current pixel and an adjacent pixel is greater than a first value and less than a second value, at least an additional point of said 3D scene is generated, in addition to a current point associated with the current pixel of the depth image, the number of additional points depending on the depth difference.