This is an old revision of the document!
3d scan of tami
camera registration (superpoint+lightglue)
coarse patchmatch mvs depth estimation
alignment of neural monocular estimate to patchmatch(ransac polynomial fitting)
dense point cloud integration
3dgs scene refinement
3dgs depth rasterization
depth kernel density outlier detection
tsdf integration
mesh compresson
voxelization
cam16_ucs color space nearest neighbors
minetest world format bullshit