BazAR: A vision based fast detection library

BazAR is a computer vision library based on feature points detection and matching. In particular, it is able to quickly detect and register known planar objects in images. Well adapted to Augmented Reality applications, it is the result of advanced computer vision research.

Watch some calibration screen shots or augmented reality videos.


Feature points detection and matching

Source code is available under the GNU General Public License. In short, if you distribute a software that uses BazAR, you have to distribute it under GPL with the source code. Another option is to contact us to purchase a commercial license.


Who wrote BazAR ?

How to get started with camera calibration and augmentation

Download and install bazar. To calibrate the camera, run the example called “singlecalib”. You will have to present a flat, rigid and textured surface to the camera, as fronto-parallel as you can. Press space, then ‘Y’. Model learning will start. After, simply move your calibration object in the field of view. After some time, calibration computation starts.

After calibration, it is possible to augment the video stream with a very basic 3d object. Just run the “Augment3D” sample.

How to get started with detection

Download the archive, install it, and run the sample program in bazar/samples/filedetector.
Under Windows, run: test.bat
Under UNIX, run : ./filedetector *jpg

Keypoint Recognition using Randomized Trees

V. Lepetit; P. Fua 

IEEE Transactions on Pattern Analysis and Machine Intelligence. 2006. Vol. 28, num. 9, p. 1465-1479. DOI : 10.1109/TPAMI.2006.188.

Randomized Trees for Real-Time Keypoint Recognition

V. Lepetit; P. Lagger; P. Fua 

2005. 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05), San Diego, CA, p. 775-781. DOI : 10.1109/CVPR.2005.288.

Links to related software