Skip to main content
. Author manuscript; available in PMC: 2022 Dec 13.
Published in final edited form as: J Mol Graph Model. 2021 Dec 21;111:108103. doi: 10.1016/j.jmgm.2021.108103

Fig. 3.

Fig. 3.

ProteinNet deep architecture for protein point cloud transformation into canonical representation. Step (1): affine transformation matrix estimation. Step (2): protein point cloud transformation using the estimated affine matrix. Step (3): similarity calculation between the original protein point cloud (the input) and its transformed point cloud. Step (4): cosine similarity loss calculation between the original input protein point cloud and its transformation; and back–propagation over the network to optimize the estimation of the affine transformation matrix.