Canonical forms for non-rigid 3D shape retrieval (source code)

Source code for two canonical form methods, which generate a standard pose for non-rigid 3D objects. There is also code for a view-based retrieval method, which can be applied to the canonical forms for non-rigid 3D shape retrieval.

The code is written in Matlab, with small portions written in C++ to interface with Matlab as mex files. The code contains README.txt files which describe how to compile and use the code, as well as a demonstration script (demo.m) for each canonical form method. The view-based retrieval method has some third-party dependencies, which are detailed in the associated README.txt file.

There are two canonical form methods implemented, which are implementations of the following two papers:

David Pickup, Xianfang Sun, Paul L. Rosin, Ralph R. Martin, "Skeleton-Based Canonical Forms for Non-Rigid 3D Shape Retrieval", Computational Visual Media, 2016.
http://dx.doi.org/10.1007/s41095-016-0045-5

Pickup, D., Sun, X., Rosin, P.L., Martin, R.R. “Euclidean-distance-based canonical forms for non-rigid 3D shape retrieval”, Pattern Recognition 2015.
http://dx.doi.org/10.1016/j.patcog.2015.02.021

Information on these and other relevant publications can be found at

http://orca.cf.ac.uk/89013/
http://orca.cf.ac.uk/71104/
http://orca.cf.ac.uk/73133/
http://orca.cf.ac.uk/89014/
http://orca.cf.ac.uk/60329/

The view-based method is my implementation of a method proposed by Zhouhui Lian in the following paper:

Z Lian, A Godil, X Sun, J Xiao, "CM-BOF: visual similarity-based 3D shape retrieval using Clock Matching and Bag-of-Features",  Machine vision and applications, 2013.