Example-based Facade Texture Synthesis and Inpainting

Dengxin Dai, Hayko Riemenschneider, Gerhard Schmitt, Luc Van Gool


There is an increased interest in the efficient creation of city models, be it virtual or as-built. We present a method for synthesizing complex, photo-realistic facade images, from a single example. After parsing the example image into its semantic components, a tiling for it is generated. Novel tilings can then be created, yielding facade textures with different dimensions or with occluded parts inpainted. A genetic algorithm guides the novel facades as well as inpainted parts to be consistent with the example, both in terms of their overall structure and their detailed textures. Promising results for multiple standard datasets, in particular for the different building styles they contain, demonstrate the potential of the method.


Facade Synthesis!

Fig1. The pipeline of our method: From a parsed example facade (a), to its grid representation (b), to a larger, synthesized grid with inferred label configuration (c), and to the synthesized facade (d). Each node in (b) has a unique label indicating its own tile and it is highlighted with a specific color.


Facade Synthesis!

Fig2. Comparison of different methods.

Facade Synthesis!

Fig3. An illustration of how the synthesis result evolves with the number of iterations, resulting in decreasing energy.

Facade Synthesis!

Fig4. Synthesis result on Paris2011 dataset.

Facade Synthesis!

Fig5. Synthesis result on FaSyn13 dataset.

Facade Synthesis!

Fig6. The method is stochastic and is able to generate different solutions from different runs.

Facade Synthesis!

Fig7. Failure cases. The artifacts in (b) are due to the confusing window textures of (a), which results in errors in the semantic labeling and lattice parsing. In (d), the global symmetric structure of (c) is lost, which is beyond the ability of our two semi-local constraints.

Facade Synthesis!

Fig8. Inpainting Results. Our method is better in keeping structures.

Facade Synthesis!

Fig9. Inpainting Results. The method can come very close to the ground truth.


Paper and Supplementary Material
D. Dai, H. Riemenschneider, G. Schmitt, and L. Van Gool. Example-based Facade Texture Synthesis. ICCV, 2013. PDF BibTex More synthesis results More inpainting results

The Poster presented in ICCV 2013.

FaSyn13 dataset (200 facade images)
This dataset contains 200 facades (200 original images + their 200 corresponding rectified, cropped images) of varying building styles, including Classicism, Dutch Romanesque, Renaissance, Modern, etc.

Other Facade/Building/City datasets

This page has been edited by Dengxin Dai