Block-Based Voxel Compression for 3D Volumetric Representations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Transmitting high-quality three and four-dimensional volumetric sequences is challenging due to their large memory footprints, particularly with mesh representations, which can be prohibitive for efficient data transmission.
Innovation Solution
A block-based end-to-end trainable geometry compression model using signed distance fields (TSDFs) for lossy compression, combined with lossy parameterization-free texture compression and lossless surface topology, employing machine-learned voxel encoding and decoding models to selectively encode and reconstruct three-dimensional volumetric representations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If mesh representation is used for volumetric data, then reconstruction quality is maintained, but memory footprint becomes prohibitively large
Solution Approach 1:
The patent divides the volumetric representation into discrete voxel blocks that can be independently encoded and transmitted. This segmentation allows selective compression of volumetric data while maintaining reconstruction quality, directly resolving the contradiction between high memory footprint and quality preservation by enabling efficient data representation through block-based processing
Solution Approach 2:
The patent transforms the representation parameters from traditional mesh-based continuous geometry to voxel-based discrete volumetric grids with associated signed distance fields. This parameter change enables lossy compression of geometry while preserving sufficient detail for high-quality reconstruction, thereby reducing memory footprint without sacrificing reconstruction quality
2Measurement precision
If high-quality three and four-dimensional sequences are transmitted, then reconstruction accuracy is maintained, but network resources and processing power increase
Solution Approach 1:
The patent extracts and separately encodes texture information from the volumetric data using parameterization-free texture compression. By separating texture encoding from geometry encoding, the system can transmit compressed geometry data with reduced network resources while maintaining reconstruction accuracy through efficient texture application during decoding
Solution Approach 2:
The patent performs preliminary compression of volumetric data into encoded voxel blocks before transmission. This preliminary encoding action reduces the data size that needs to be transmitted over the network, thereby conserving network resources and processing power while preserving the ability to reconstruct high-quality sequences at the receiving end
Data Source
AI summary
Systems and methods are directed to encoding and/or decoding of the textures/geometry of a three-dimensional volumetric representation. An encoding computing system can obtain voxel blocks from a three-dimensional volumetric representation of an object. The encoding computing system can encode voxel blocks with a machine-learned voxel encoding model to obtain encoded voxel blocks. The encoding computing system can decode the encoded voxel blocks with a machine-learned voxel decoding model to obtain reconstructed voxel blocks. The encoding computing system can generate a reconstructed mesh representation of the object based at least in part on the one or more reconstructed voxel blocks. The encoding computing system can encode textures associated with the voxel blocks according to an encoding scheme and based at least in part on the reconstructed mesh representation of the object to obtain encoded textures.