A manual about how to use "Evaluation.m" in Matlab and "psbPlot.exe" to evaluate the retrieval accuracy of your method tested on the SHREC'15 Track: Non-rigid 3D Shape Retrieval (the code is the same as the one used in SHREC'11 Non-rigid track):

1. Rename your distance matrix file as "result.matrix";

2. Put the classification file and your distance matrix in the same folder, for example "H:\LIAN\CODE\Matlab\";

3. In the 30th line of "Evaluation.m", assign the variable "filepath" a string of the folder's file path, for instance, " filePath = 'H:\LIAN\CODE\Matlab\' ";

4. Run "Evaluation.m".


Check your result in the file "result.txt". The other file generated by this code is "PSBresult.matrix", which can be evaluated using the source codes provided by the Princeton Shape Benchmark
http://shape.cs.princeton.edu/benchmark/util.html

For example, Precision-recall plots can be obtained using the utility "psbPlot.exe" by typing the following command:

-- psbPlot test.cla PSBresult.matrix


*************************************************************
Please cite the following papers:

@inproceedings{3DOR11:79-88:2011,
crossref = {3DOR11-proc},
author = {Z. Lian and A. Godil and B. Bustos and M. Daoudi and J. Hermans and S. Kawamura and Y. Kurita and G. Lavou and H. V. Nguyen and R. Ohbuchi and Y. Ohkita and Y. Ohishi and F. Porikli and M. Reuter and I. Sipiran and D. Smeets and P. Suetens and H. Tabia and D. Vandermeulen },
title = {{SHREC '11 Track: Shape Retrieval on Non-rigid 3D Watertight Meshes}},
pages = {79-88},
URL = {http://diglib.eg.org/EG/DL/WS/3DOR/3DOR11/079-088.pdf},
}

@inproceedings{3DOR15:Lian:2015,
crossref = {3DOR15-proc},
author = {Z. Lian and et al.},
title = {{SHREC '15 Track: Non-rigid 3D Shape Retrieval}},
}
*************************************************************


