ASL Object Recognition ToolKit

Snapshot of ORTK ORTK is a multi-platform implementation of the object detection system presented by David G. Lowe [1]. The system is still under development but so far it includes the main features of the original system plus a multi-view system to recognize objects from widely separated points of view. The system has been tested in [2].
The implementation is done in C++ using the QT and the OpenCV libraries.

This software has been released under the General Public License (GPL)

Screenshots

Download

Project and sample images:
ortk.tar.gz (7.6 MB)
ortk.zip (7.6 MB)

Project without sample images:
ortk-noimages.tar.gz (84.8 KB)
ortk-noimages.zip (95.7 KB)

Readme file

Project Team

Laurent Winkler
Cyril Schmidiger
Davide Scaramuzza
Arnau Ramisa (aramisa@iiia.csic.es)
Shrihari Vasudevan (shrihari.vasudevan@ieee.org)

ASL object recognition database

  1. 9 objects (3 textured, 3 uniformly textured and 3 with no texture)
  2. 36 test images containing random subsets of these objects
  3. Ground Truth
  4. Notes about dataset: Challenging dataset + Blurred images + changing illumination
  5. Reference Publication : Ramisa, Vasudevan et al @ ICVS 2008 : please refer [2]

References

[1] David G. Lowe, "Distinctive image features from scale-invariant keypoints", International Journal of Computer Vision, 60, 2 (2004), pp. 91-110.

[2] Arnau Ramisa, Shrihari Vasudevan, Davide Scaramuzza, Ramon Lopez de Mantaras and Roland Siegwart, " A tale of two object recognition methods for mobile robots", In the proceedings of the International Conference on Computer Vision Systems (ICVS) 2008, Santorini, Greece.