Ray-ONet: Efficient 3D Reconstruction From A Single RGB Image

Wenjing Bian, Zirui Wang, Kejie Li, Victor Adrian Prisacariu
Active Vision Lab, University of Oxford

Arxiv Bibtex Code

Abstract

We propose Ray-ONet to reconstruct detailed 3D models from monocular images efficiently. By predicting a series of occupancy probabilities along a ray that is back-projected from a pixel in the camera coordinate, our method Ray-ONet improves the reconstruction accuracy in comparison with Occupancy Networks (ONet), while reducing the network inference complexity to O(\(N^2\)). As a result, Ray-ONet achieves state-of-the-art performance on the ShapeNet benchmark with more than 20\(\times\) speed-up at \(128^3\) resolution and maintains a similar memory footprint during inference.

Visualisation

We show the 3D reconstruction results on three datasets: ShapeNet, Online Products, Pix3D. Our model is trained with ShapeNet objects in 13 categories.

BibTeX

  @inproceedings{bian2021rayonet,
	title={Ray-ONet: Efficient 3D Reconstruction From A Single RGB Image}, 
	author={Wenjing Bian and Zirui Wang and Kejie Li and Victor Adrian Prisacariu},
	booktitle={BMVC},
	year={2021}
   }