Point Cloud ROI Tile Packing for HEVC Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current compression technologies for point clouds in 3D video coding fail to efficiently compress 3D geometry and attributes, particularly in immersive applications like virtual reality and autonomous driving, where lossy and lossless compression, scalable coding, and random access are required, and existing methods do not guarantee that 3D regions of interest are projected into fewer tiles efficiently.
Innovation Solution
The method involves separating point clouds into region-of-interest (ROI) and non-ROI clouds, generating and packing patches into a 2D grid, ensuring ROI patches are placed in as few tiles as possible, and using High Efficiency Video Coding (HEVC) tiles to prioritize ROI compression, allowing for efficient content-aware point cloud compression that maintains high quality for ROIs while optimizing bit rate.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If point clouds are compressed using existing compression technologies, then compression is achieved, but ROI quality is not prioritized and bit rate efficiency is insufficient
Solution Approach 1:
The point cloud is segmented into multiple patches, with ROI patches identified and separated from non-ROI patches. This segmentation allows differential compression where ROI patches receive higher quality encoding while non-ROI patches use more aggressive compression, resolving the contradiction between ROI quality and bit rate efficiency.
Solution Approach 2:
Different compression qualities are applied to different regions of the point cloud. ROI patches are encoded with higher quality parameters while non-ROI patches use lower quality parameters. This local quality differentiation maintains high ROI quality while improving overall bit rate efficiency.
2Manufacturing precision
If point clouds are compressed with high quality for all regions, then overall quality is maintained, but bit rate consumption increases significantly
Solution Approach 1:
The patent applies different compression qualities to different spatial regions. ROI patches maintain high quality while non-ROI patches use lower quality encoding. This resolves the contradiction by maintaining acceptable overall quality while significantly reducing bit rate consumption through selective quality allocation.
Solution Approach 2:
High quality compression is applied only to the necessary ROI portions rather than the entire point cloud. This partial action approach maintains quality where needed while reducing overall bit rate consumption by applying aggressive compression to non-ROI regions.
3Productivity
If ROI patches are mapped to multiple tiles, then spatial distribution is optimized, but compression efficiency decreases
Solution Approach 1:
Multiple ROI patches that would otherwise be distributed across multiple tiles are merged and packed into a single tile. This consolidation improves compression efficiency by reducing the number of tile boundaries and overhead, while maintaining spatial distribution benefits through the packing arrangement within the unified tile structure.
Data Source
AI summary
A method includes receiving a data cloud including a plurality of data points. The method further includes identifying each data point including a region-of-interest (ROI) and dividing the data cloud into a ROI cloud and one or more non-ROI clouds. The method includes performing a patch generation process on the ROI cloud, the patch generation process including generating a ROI patch from each data point including the ROI. The method includes performing a patch packing process on the ROI cloud, the patch packing process including: (i) mapping each ROI patch to a two dimensional (2D) map, (ii) determining whether at least two ROI patches from the plurality of ROI patches are located in more than one tile of the map, and (iii) in response to the determination that at least two ROI patches are located in more than one tile, moving each of the ROI patches to a tile.