Theia

Theia is a computer vision library developed by Chris Sweeney aimed at providing efficient and reliable algorithms for Structure from Motion (SfM). The goal of this library is to provide researchers with an out of the box tool for multi-view reconstruction that can be easily extended. Many common algorithms for pose, feature detection and description, matching, and reconstruction have been implemented. All contain simple interfaces, limited dependencies, and extensive documentation.

Documentation

To use Theia, simply add the following line to your program after you build and link the library:

#include <theia/theia.h>

We attempt to provide sufficient documentation but often further documentation can be found in the source code itself. You will likely find the API documentation useful as well. Additionally, (nearly) every file is covered by a unit test that can be viewed as an example use case of the various methods and classes in Theia. If you have looked at the documentation, the tutorials, the source code, and the unit tests and still have confusion please email the Theia mailing list

Finally, it should be noted that all the code in Theia is under the namespace theia, so you will have to reference that namespace in order to use functions from this library.

Performance

Theia achieves state-of-the-art SfM performance on large-scale datasets. Efficiency and robustness is a key component of the library. You can see the latest performance benchmarks for small and large-scale datasets on the Performance page.

Citation

If you use Theia for an academic publication, please cite this manual. e.g.,

@manual{theia-manual,
        Author = {Chris Sweeney},
        Title = {Theia Multiview Geometry Library: Tutorial \& Reference},
        Organization = {University of California Santa Barbara.}
}

Acknowledgements

Theia was originally developed to provide a centralized code base to the Four Eyes Lab at UC Santa Barbara, but has since been expanded to an open-source project for the vision community.

The core of the original library is written by Chris Sweeney. Funding for Theia was provided by his advisors Tobias Hollerer and Matthew Turk and NSF Graduate Research Fellowship Grant DGE-1144085.