Manifold Learning. Figure 1: Misaligned dataset is learned in the space of 3D shapes quotiented by rotations, thus the quotient autoencoder focuses on.
確定! 回上一頁