Factored Facade Acquisition using Symmetric Line Arrangements

Duygu Ceylan, Niloy J. Mitra, Hao Li, Thibaut Weise, Mark Pauly

Eurographics 2012

[pdf] [video] [bibtex] [fast forward] [slides]

Factored Facade Acquisition using Symmetric Line Arrangements

We introduce a novel framework for image-based 3D reconstruction of urban buildings based on symmetry priors. Starting from image-level edges, we generate a sparse and approximate set of consistent 3D lines. These lines are then used to simultaneously detect symmetric line arrangements while refining the estimated 3D model. Operating both on 2D image data and intermediate 3D feature representations, we perform iterative feature consolidation and effective outlier pruning, thus eliminating reconstruction artifacts arising from ambiguous or wrong stereo matches. We exploit non-local coherence of symmetric elements to generate precise model reconstructions, even in the presence of a significant amount of outlier image-edges arising from reflections, shadows, outlier objects, etc. We evaluate our algorithm on several challenging test scenarios, both synthetic and real. Beyond reconstruction, the extracted symmetry patterns are useful towards interactive and intuitive model manipulations.

Data Sets

For each data set, the input images, the per-image camera projection matrices, and our final reconstructions are provided. For the synthetic example, we provide the ground-truth camera parameters extracted from Maya together with the ground-truth model. The calibration information is obtained by running the open-source structure-from-motion code Bundler with default parameters. We use the output format compatible with PMVS


You can download the complete data set here.

SoHo Data Set (13 images)

soho-images.zip

soho-cameras.zip

soho-model.zip

Synthetic Data Set (24 images)

synthetic-images.zip

synthetic-cam.zip (Camera parameters extracted from Maya)

synthetic-model.zip

groundTruthModel.zip (Ground-truth model)