README.md

metadata

license: apache-2.0
tags:
  - Graphics
  - 3D_reconstruction
  - Fisheye
pretty_name: FullCircle

FullCircle Dataset

Dataset for FullCircle: Effortless 3D Reconstruction from Casual 360° Captures.

Summary

Per-scene assets for casual 360° captures: paired fisheye views, masks, equirectangular omni frames, and a COLMAP sparse model. Use with the official code for training or evaluation.

Structure

Each scene is organized as follows:

<scene_name>/
├── images/          # fisheye frames (camera1/, camera2/)
├── masks/           # fisheye masks (camera1/, camera2/)
├── omni/images/     # 360° omnidirectional frames
└── sparse/0/        # COLMAP reconstruction

Usage

Clone theialab/fullcircle, put each scene under data/<scene_name>/ (matching the tree above), then follow that repo’s training docs.

Citation

@article{foroutan2026fullcircle,
  title   = {FullCircle: Effortless 3D Reconstruction from Casual 360° Captures},
  author  = {Foroutan, Yalda and Oztas, Ipek and Rebain, Daniel and Dundar, Aysegul and Yi, Kwang Moo and Goli, Lily and Tagliasacchi, Andrea},
  journal = {arXiv preprint arXiv:2603.22572},
  year    = {2026}
}