CSfM: Community-based Structure from Motion

TL;DR

The paper introduces CSfM, a community-based Structure from Motion method that improves robustness and efficiency by partitioning the epipolar graph and merging reconstructions, outperforming many existing approaches.

Contribution

It proposes a novel community-based SfM framework with parallel reconstruction and a global similarity averaging method for improved robustness and efficiency.

Findings

Outperforms many state-of-the-art global SfM methods in efficiency.

Achieves similar or better accuracy and robustness compared to incremental SfM.

Uses convex L1 optimization for merging reconstructions.

Abstract

Structure-from-Motion approaches could be broadly divided into two classes: incremental and global. While incremental manner is robust to outliers, it suffers from error accumulation and heavy computation load. The global manner has the advantage of simultaneously estimating all camera poses, but it is usually sensitive to epipolar geometry outliers. In this paper, we propose an adaptive community-based SfM (CSfM) method which takes both robustness and efficiency into consideration. First, the epipolar geometry graph is partitioned into separate communities. Then, the reconstruction problem is solved for each community in parallel. Finally, the reconstruction results are merged by a novel global similarity averaging method, which solves three convex $L1$ optimization problems. Experimental results show that our method performs better than many of the state-of-the-art global SfM approaches in terms of computational efficiency, while achieves similar or better reconstruction accuracy and robustness than many of the state-of-the-art incremental SfM approaches.