We learn a self-supervised, single-view 3D reconstruction model that predicts the 3D mesh
shape, texture and camera pose of a target object with a collection of 2D images and
silhouettes. The proposed method does not necessitate 3D supervision, manually annotated
keypoints, multi-view images of an object or a prior 3D template. The key insight of our work
is that objects can be represented as a collection of deformable parts, and each part is
semantically coherent across different instances of the same category (eg, wings on birds …

US11238650B2 - Self-supervised single-view 3D reconstruction via semantic consistency -
Google Patents US11238650B2 - Self-supervised single-view 3D reconstruction via semantic
consistency - Google Patents Self-supervised single-view 3D reconstruction via semantic
consistency Download PDF Info Publication number US11238650B2 US11238650B2
US16/849,962 US202016849962A US11238650B2 US 11238650 B2 US11238650 B2 US
11238650B2 US 202016849962 A US202016849962 A US 202016849962A US 11238650 B2 …