FIGTree: Fast Improved Gauss Transform with Tree Data Structure  

A library for fast computation of Gauss transforms in multiple dimensions, using the Improved Fast Gauss Transform and Approximate Nearest Neighbor searching. The nearest neighbor searching is performed using the ANN library, available at http://www.cs.umd.edu/~mount/ANN/. This software allows for efficient computation of probabilities by Kernel Density Estimation (KDE), and can reduce complexity of algorithms commonly used in Computer Vision, Machine Learning, etc, that must evaluate the Gauss transform.

Related Publications:

The publication describing the newest improvements in the code is the NIPS 2008 paper by Morariu et al (see Vlad Morariu's site). Previous publications related to this approach are provided on Vikas Raykar's page.
NOTE: The version of the code currently available on SourceForge does not include all automatic method and parameter selection approaches used for automatic tuning described in the NIPS 2008 paper (i.e. the method of evaluation still needs to be chosen manually by the user). The code will be updated soon with these changes that will allow FIGTree to be used as a black box.

Authors:

Vikas Raykar and Changjiang Yang were the initial authors of previous versions of IFGT and FIGTree. The C/C++ interface and other modifications to Vikas Raykar's version of the code were made by Vlad Morariu. The authors worked under the supervision of Professor Ramani Duraiswami and Professor Larry Davis, at the University of Maryland.

License:

This code extends Vikas Raykar's version of the IFGT code, which was provided under the GNU Lesser General Public License (LGPL). As a result, the FIGTree library is also released under the LGPL.

Download:

Download (from SourceForge)