Shape, Pose, and Appearance from a Single Image via Bootstrapped Radiance Field Inversion#
Authors: Dario Pavllo, David Joseph Tan, Marie-Julie Rakotosaona, Federico Tombari
Affiliations: Google, ETH Zurich, TU Munich
Summary#
The authors introduced a principled end-to-end reconstruction framework for natural images, where accurate ground-truth poses are not available. The approach recovers an SDF-parameterized 3D shape, pose, and appearance from a single image of an object, with exploiting multiple views during training. The hybrid GAN inversion technique includes an encoder that produces a first guess of the solution, which is then refined by optimization.