This is a dataset derived from the datasets available at http://cvlabwww.epfl.ch/~strecha/multiview/denseMVS.html and
http://www.cs.unc.edu/~jheinly/feature-evaluation/datasets.html, containing images with different geometric and photometric transformations like rotation, scale, viewpoint, image blur, JPEG compression and illumination of structured and textured scenes. These original datasets were presented by C. Strecha and W. von Hansen and L. Van Gool and P. Fua and U. Thoennessen “On Benchmarking Camera Calibration and Multi-View Stereo for High Resolution Imagery” and J. Heinly, E. Dunn and J. M. Frahm “Comparative Evaluation of Binary Features”
How to obtain this dataset
This Database can be used for non-commercial research provided you acknowledge the source and also original source from which the images were derived:
C. Strecha and W. von Hansen and L. Van Gool and P. Fua and U. Thoennessen “On Benchmarking Camera Calibration and Multi-View Stereo for High Resolution Imagery”, Computer Vision and Pattern Recognition, 2008. CVPR 2008. IEEE Conference on , vol., no., pp.1,8, 23-28 June 2008.
J. Heinly, E. Dunn and J. M. Frahm “Comparative Evaluation of Binary Features”, Computer Vision – ECCV 2012, Springer Berlin Heidelberg, 2012.
I. O. de Oliveira and E. Todt, “Improved ORB Feature Detector”, submitted to SIBGRAPI 2014.
Before you can download the database we ask you to register by sending us an email (vri AT inf.ufpr.br) so we are aware of who is using our data. Please include in your email your Last name, First name, Organization and email. Once you have registered you will receive a link to download the database.