Point Cloud Coding Using RAHT and Spatial Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current point cloud coding methods, such as those used in the geometry-based point cloud compression (G-PCC) framework, suffer from low encoding efficiency due to the lack of consideration for spatial conditions of current nodes in the point cloud during upsampled prediction.
Innovation Solution
The proposed method improves encoding efficiency by determining reconstructed attribute information of occupied neighboring nodes and same-level neighboring child nodes, using this information to calculate an attribute prediction value, and then performing region adaptive hierarchal transform (RAHT) on this value to obtain transformed coefficients for reconstructing attribute information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If upsampled prediction is used to remove spatial redundancy, then compression performance is improved, but encoding efficiency deteriorates because spatial conditions of current nodes are not considered
Solution Approach 1:
The patent applies local quality by considering the spatial conditions of current nodes specifically in the prediction process. Instead of using a uniform prediction approach, the method determines whether to perform upsampled prediction based on the spatial characteristics of each current node, thereby optimizing the balance between compression performance and encoding efficiency for different local regions.
2Productivity
If same-level neighboring child nodes are introduced for prediction, then coding efficiency is improved, but device complexity increases
Solution Approach 1:
The patent implements preliminary action by pre-determining the spatial conditions of current nodes and pre-identifying potential neighboring nodes before the actual prediction process. This allows the system to prepare the necessary prediction data and structures in advance, reducing the computational burden during encoding while still benefiting from the enhanced prediction accuracy provided by same-level neighboring child nodes.
Data Source
AI summary
A point cloud decoding method is applied to a decoder and includes the following. Reconstructed attribute information of occupied neighbouring nodes of a current node and reconstructed attribute information of occupied same-level neighbouring child nodes of a child node of the current node are determined. An attribute prediction value of the child node of the current node is determined according to the reconstructed attribute information of the occupied neighbouring nodes and the reconstructed attribute information of the occupied same-level neighbouring child nodes of the child node of the current node. A first transformed coefficient is determined by performing region adaptive hierarchal transform (RAHT) on the attribute prediction value, and reconstructed attribute information of the child node of the current node is determined according to the first transformed coefficient.