ShapeFromBlur

[NeurIPS 2021] Shape from Blur: Recovering Textured 3D Shape and Motion of Fast Moving Objects

YouTube | arXiv

<img src="examples/imgs/aerobie.gif" height="160">       <img src="examples/imgs/football.gif" height="160">       <img src="examples/imgs/vol.gif" height="160">       <img src="examples/imgs/key__new.gif" height="160">

Prerequisites

Kaolin is available here: https://github.com/NVIDIAGameWorks/kaolin

Running

The code can be easily run by:

python optimize.py

Running with your inputs:

python optimize.py --im examples/vol_im.png --bgr examples/vol_bgr.png
python optimize.py --im examples/aerobie_im.png --bgr examples/aerobie_bgr.png
python optimize.py --im examples/pen_im.png --bgr examples/pen_bgr.png

The results will be written to the output folder.

Reference

If you use this repository, please cite the following publication:

@inproceedings{sfb,
  title = {Shape from Blur: Recovering Textured 3D Shape and Motion of Fast Moving Objects},
  author = {Denys Rozumnyi and Martin R. Oswald and Vittorio Ferrari and Marc Pollefeys},
  booktitle = {NeurIPS},
  month = {Dec},
  year = {2021}
}