Основные понятия
Efficient lossless compression of point cloud geometry and color attributes using sparse tensor-based neural networks.
Аннотация
The article introduces a novel method for lossless compression of point cloud data, focusing on geometry and color attributes. It utilizes sparse tensor-based deep neural networks to achieve efficient compression. The method outperforms existing standards like MPEG G-PCC, achieving significant reductions in total bitrate while maintaining accuracy.
Structure:
- Introduction to Point Cloud Data Importance
- Existing Standards: V-PCC and G-PCC by MPEG
- Proposed Method: CNeT for Lossless Compression
- Network Architecture and Context Modeling
- Experimental Results and Comparison with Existing Methods
Статистика
Compared with MPEG's G-PCC, the proposed method achieves a 22.6% reduction in total bitrate.
The method shows a 49% rate reduction on geometry components and an 18.3% reduction on color attribute components.
Цитаты
"Our method represents a point cloud with both occupancy feature and three attribute features at different bit depths in a unified sparse representation."
"Compared with the state-of-the-art lossless point cloud compression method from Moving Picture Experts Group (MPEG), our method achieves 22.6% reduction in total bitrate."