Abstract: We present a deep learning pipeline that leverages network self-prior to recover a full 3D model consisting of both a triangular mesh and a texture map from the colored 3D point cloud.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results