This paper proposes an extreme Structure from Motion (SfM) algorithm for residential indoor panoramas that have little to no visual overlaps. Only a single panorama is present in a room for many cases, making the task infeasible for existing SfM algorithms. Our idea is to learn to evaluate the realism of room/door/window arrangements in the top-down semantic space. After using heuristics to enumerate possible arrangements based on door detections, we evaluate their realism scores, pick the most realistic arrangement, and return the corresponding camera poses. We evaluate the proposed approach on a dataset of 1029 panorama images with 286 houses. Our qualitative and quantitative evaluations show that an existing SfM approach completely fails for most of the houses. The proposed approach achieves the mean positional error of less than 1.0 meter for 47% of the houses and even 78% when considering the top five reconstructions.



Code and Data

Please visit the GitHub page of our project for more details.


author = {Shabani, Mohammad Amin and Song, Weilian and 
        Odamaki, Makoto and Fujiki, Hirochika and Furukawa, Yasutaka},
title = {Extreme Structure from Motion for Indoor Panoramas without Visual Overlaps},
booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
month = {October},
year = {2021},
url = {https://aminshabani.github.io/publications/extreme_sfm/pdfs/iccv2021_2088.pdf}