A New Method for Single Shot Scene Reconstruction from RGB Images using Implicit Representations

Systems and methods described herein relate to reconstructing a scene in three dimensions from a two-dimensional image. One embodiment processes an image using a detection transformer to detect an object in the scene and to generate a NOCS map of the object and a background depth map; uses MLPs to relate the object to a differentiable database of object priors (PriorDB); recovers, from the NOCS map, a partial 3D object shape; estimates an initial object pose; fits a PriorDB object prior to align in geometry and appearance with the partial 3D shape to produce a complete shape and refines the initial pose estimate; generates an editable and re-renderable 3D scene reconstruction based, at least in part, on the complete shape, the refined pose estimate, and the depth map; and controls the operation of a robot based, at least in part, on the editable and re-renderable 3D scene reconstruction.

Researchers

Fredo Durand (he/him/his) / Jiajun Wu / Dennis Park / Rares Ambrus / Adrien Gaidon / Vitor Guizilini / Wadim Kehl / Sergey Zakharov / Vincent Sitzmann / Joshua B Tenenbaum

Departments: Dept of Electrical Engineering & Computer Science, Department of Brain and Cognitive Sciences
Technology Areas: Artificial Intelligence (AI) and Machine Learning (ML) / Computer Science: Bioinformatics / Sensing & Imaging: Optical Sensing

  • systems and methods for reconstructing a scene in three dimensions from a two-dimensional image
    United States of America | Granted | 11,887,248

License this technology

Interested in this technology? Connect with our experienced licensing team to initiate the process.

Sign up for technology updates

Sign up now to receive the latest updates on cutting-edge technologies and innovations.