Surface Context (original) (raw)

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

Software is available here.

Dataset

Data is available here.

Results

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)

Publications

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