Radial Distorted Images : RD sets

This a support page with data sets for evaluation of keypoints and descriptors computed for radial distorted images. We provide two different data set to perform both repetability of detection and also structure from motion. This work was fully financed by the Project ArthroNav.

The data provided is organized as follows:

  • Planar scenes : related by an homography between 78 pairs of images. This images are used to compute detection repetibility and retrieval performance automatically using the provided projective transformation between planes (click on the planar sequence image to download)
  • Depth variations scenes : 7 images pairs of images for each amount of distortion provided. Useful to validate keypoint usability in motion and structure recovery. (click on the SFM sequence image to download)

This data sets were acquired using 3 types of lens, with low, moderate and high distortion. The data is fully calibrated with the amazing EasyCamCalib toolbox done by my friend Rui Melo and my advisor Prof. Barreto for calibration of images with radial distortion from a single grid.


Related Papers

M. Lourenco, J.P. Barreto, and F. Vasconcelos, "sRD-SIFT Keypoint Detection and Matching in Images with Radial Distortion", " IEEE Transactions on Robotics 2012.

M. Lourenco, J.P. Barreto, and A. Malti, "Feature Detection and Matching in Images with Radial Distortion," IEEE International Conference on Robotics and Automation, 2010. (Best Student Paper Award Finalist)