A Method Based on Curvature and Hierarchical Strategy for Dynamic Point Cloud Compression in Augmented and Virtual...

Area of Science:

• Computer Vision
• 3D Data Processing
• Signal Compression

Background:

• Dynamic point clouds are crucial for immersive applications.
• Existing compression methods often ignore point cloud characteristics, leading to slow encoder runtimes.
• Efficient compression is needed for real-world dynamic point cloud applications.

Purpose of the Study:

• To develop an improved compression method for dynamic point clouds.
• To address the limitations of exhaustive neighborhood search in current methods.
• To enhance both compression performance and encoder speed.

Main Methods:

• A novel dynamic point cloud compression approach utilizing curvature estimation and a hierarchical strategy.
• Initial segmentation based on normal vector similarity.
• Curvature-based hierarchical refinement of voxel point clouds, distinguishing high- and low-curvature regions.
• Integration of image generation and video compression techniques for de-redundancy.

Main Results:

• The proposed method demonstrates superior compression performance compared to traditional video-based dynamic point cloud compression.
• Achieved a faster runtime for the encoder.
• Effectively optimized initial clusters through hierarchical refinement.

Conclusions:

• The curvature-based hierarchical strategy offers a more efficient approach to dynamic point cloud compression.
• This method meets the demands of real-world scenarios by balancing compression ratio and speed.
• Significant improvements in both compression efficiency and processing speed were validated.