| | | |

3D Technologies

3D from Paintings and Old Photos

Examples of converting paintings and photos to 3D

	Hooch
	Steenwyck
	Saenredam
	Old City Hall
	From the movie “2001: A Space Odyssey”

In the 3D from Paintings and Old Photos research, advanced representation and visualization methods are used in developing software that converts data into 3D models. The new technology developed advances applications in such fields as computer graphics, animation and special effects. This research was displayed in the short movie “Portals”, one of the highlights of the prestigious SIGGRAPH 2002 conference on computer graphics and interactive technology. In “Portals”, viewers find themselves travelling through monuments and flying into Renaissance paintings. In the following examples, you can see the first steps of how a painting or digital image can be rendered into 3D.

The 3D from Paintings and Old Photos project started in 2000, and is expected to end in 2005.

Research context of the project

Rationale: Traditional image-based 3D reconstruction methods use multiple images to extract 3D geometry. However, it is not always possible to obtain such images, such as when reconstructing destroyed structures using existing photographs or paintings with proper perspective.

Existing Techniques: Methods for 3D reconstruction from a single image do exist. They use blocks or known shapes to reconstruct from a single image. They also solve for only a simple camera model using vanishing lines and vanishing points.

The Approach: The approach does not need models of the objects nor known internal camera calibration parameters. It also does not use vanishing lines or vanishing points, which may be unavailable or hard to extract. We use several types of constraints: point/coordinate constraints, surface constraints, and topological constraints. We solve first for camera internal and external parameters using one set of constraints, then we use additional constraints to obtain 3D coordinates for reconstruction. The camera parameters and 3D coordinates are computed from photogrammetric bundle adjustment; a simultaneous triangulation of all data and the constraints.

Related NRC-IIT Publications

Research Contact

Dr. Sabry El-Hakim
Research Officer
Visual Information Technology
NRC Institute for Information Technology
1200 Montreal Road
Building M-50, Room 358
Ottawa, ON K1A 0R6
Telephone: +1 (613) 991-6381
Fax: +1 (613) 952-0215
E-mail: Sabry El-Hakim

Business Contact

Charles Gauthier
Business Development Officer
Business Development Office, NCR
NRC Institute for Information Technology
1200 Montreal Road
Ottawa, ON K1A 0R6
Telephone: +1 (613) 993-2491
Fax: +1 (613) 952-7998
E-mail: Charles Gauthier