UALG Vision Lab Toolbox
This is a programming toolbox for computer vision applications, produced by Vision Lab at UAlg (Universidade do Algarve).
This toolbox implements keypoint, line, and edge descriptors for fast object detection in images with complex background. Object detection is interfaced to Dynamic Neural Field (DFT) framework Cedar.
This release of the toolbox was produced for the Neural Dynamics project within the 7:th Framework Programme of EU (grant #270247). The repository is maintained by Dr. Erik Billing, Interaction Lab, University of Skövde, Sweden.
All code is licensed under LGPL v2.1 or any later version
- CMake 2.8+
- OpenCV 2.3+
- OpenMP (optional)
- Boost (optional, CUDA and CL only)
- Cedar (optional, http://cedar.ini.rub.de)
keypoints/ - keypoints and lines and edges descriptors/ - descriptors for characterising keypoints symmetry/ - symmetry operator disparity/ - disparity estimation evaluation/ - code for easy processing of many benchmark datasets objrec/ - object recognition work cuda/ - CUDA-based keypoints and lines/edges (not built by default) opencl/ - CL-based keypoints and lines/edges (not built by default) cedar/ - Interface to Cedar
Assuming that you have Mercurial installed on your system, and that you would like your source downloaded to ~/ObjectDetection, do:
hg clone https://bitbucket.org/interactionlab/objectdetection ~/ObjectDetection
Compilation of main library:
cd ~/ObjectDetection cp CMakeLists.txt.template CMakeLists.txt mkdir build cd build cmake .. make make doc
Compilation of Cedar plgin:
cd ~/ObjectDetection/cedar mkdir build cd build cmake .. make
- Kasim Terzić, Universidade do Algarve, Portugal