As mentioned in the "Limitations and Future Work" section of the paper, "Our current approach
requires a well-reconstructed point cloud as initialization."
I'm wondering about how to get this well-reconstructed point cloud.
Looking forward to any answers~
As mentioned in the "Limitations and Future Work" section of the paper, "Our current approach
requires a well-reconstructed point cloud as initialization."
I'm wondering about how to get this well-reconstructed point cloud.
Looking forward to any answers~