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. To tackle these problems, global

manner simultaneously estimates all camera poses, but is usu-

ally 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 parted into independent

communities. Then, the reconstruction problem is solved for

each community in parallel. Finally, a global similarity aver-

aging method is proposed to merge the reconstruction results

by solving three convex L1 optimization problems. Experi-

mental results demonstrate our method performs better than

many of the global SfM approaches in terms of efficiency,

while achieves similar or better reconstruction accuracy and

robustness than many of the state-of-the-art incremental SfM

approaches.