Camouflaging an Object from Many Viewpoints

Andrew Owens     Connelly Barnes     Alex Flint     Hanumant Singh     William Freeman    

CVPR 2014 (Oral presentation)

We address the problem of camouflaging a 3D object from the many viewpoints that one might see it from. Given photographs of an object's surroundings, we produce a surface texture that will make the object difficult for a human to detect. To do this, we introduce several background matching algorithms that attempt to make the object look like whatever is behind it. Of course, it is impossible to exactly match the background from every possible viewpoint. Thus our models are forced to make trade-offs between different perceptual factors, such as the conspicuousness of the occlusion boundaries and the amount of texture distortion. We use experiments with human subjects to evaluate the effectiveness of these models for the task of camouflaging a cube, finding that they significantly outperform näive strategies.
Paper (high-res version)
Slides (pdf, key)
Data (5GB)

Try to find the hidden box in a camouflage game! We'll use your input in our studies.

Results from selected scenes. This video shows virtual 3D boxes that have been camouflaged using our algorithms. These algorithms produce patterns that we "paint" onto the boxes, making them hard to see from many viewpoints. We show each box from the viewpoints that it has been camouflaged from. For each scene, we show the result from the algorithm with the best score (as measured by confusion rate).

Press:  WiredBoston GlobeFast Company Co.ExistThe Atlantic CitylabMIT News

  title={Camouflaging an Object from Many Viewpoints},
  author={Andrew Owens and Connelly Barnes and Alex Flint and Hanumant Singh and William Freeman},