Weighted Centroid Positioning for Point Cloud Triangle Voxelization

The patent applies parameter changes by computing optimized centroid positions through weighted averages based on triangle areas or vertex distances. Instead of using simple unweighted centroids, the method calculates weights for each vertex based on geometric parameters (triangle area or distance to centroid), then uses these weights to compute an optimized centroid position that minimizes rendering distortion. This transforms the centroid calculation from a simple arithmetic mean to a weighted mean that accounts for geometric relationships.

The patent applies preliminary action by pre-computing optimized centroid positions and associated weights before the actual compression and encoding process. The centroid optimization is performed in advance using the vertex positions and geometric relationships, so that when compression is applied, the pre-optimized centroids serve as reference values. This allows the compression algorithm to work with geometrically accurate representative points, reducing the need for aggressive compression that would introduce artifacts.