Surface Context

We extend our framework from Automatic Photo Pop-up by subclassifying vertical regions into planar (facing left, center, or right) and non- planar (porous and solid). We also provide extensive quantitative evaluation and demonstrate the usefulness of the geometric labels as context for object detection.

This work is part of an on-going effort in Geometrically Coherent Image Interpretation. In our SIGGRAPH'05 paper Automatic Photo Pop-up, we show how to construct simple "pop up" 3D models from a single image. In our CVPR'06 paper, Putting Objects in Perspective, we show how 3D reasoning can be used to aid in object detection.


Software is available here.


Data is available here.


See README.txt for description. Note that these results contain images that were obtained using Google image search. The original content providers retain copyrights.
Geometric Labeling Results
Object Detection Examples
3D Model Example (Cortona 4.2 recommended)


D. Hoiem, A.A. Efros, and M. Hebert, "Geometric Context from a Single Image", ICCV 2005. pdf

D. Hoiem, A.A. Efros, and M. Hebert, "Recovering Surface Layout from an Image", IJCV, Vol. 75, No. 1, October 2007. pdf


Some of this material is based upon work supported by the National Science Foundation under CAREER Grant No. ISS-0546547. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation. This work is also partially supported by a Microsoft Research Fellowship awarded in 2006.