8 Mar 2011
STDetect is an implementation of a Spectrogram Track Detection algorithm which was developed at the University of York by Dr. Tom Lampert and Dr. Simon O'Keefe. The aim of releasing this software is to make the algorithm available to the research community to verify published results, use as a comparison method, and, hopefully, to progress the research field.
The algorithm is based upon the active contour algorithm and uses machine learning techniques to calculate filters from the data, these form the basis of the active contour's external energy function.
Additional low-level feature detection methods, that were developed during the research, are also included (namely the Bar detector and the PCA detector, see the associated research papers here for more details).
Several Matlab scripts are included which will recreate the results presented in the thesis and research papers. The data used for this can be downloaded here