This repository contains an implementation of the Sparse Iterative Closest Point. It was implemented for the course Nuage de Point et Modélisation at Master MVA. My report regarding this work can be found here.
The dependencies are header-only and are all included in the ext directory. As a consequence, there is nothing to do. For the record, here is the list of dependencies :
When the program has been built thanks to CMake and Make, just run ./icpSparse -h in order to see the help.
Other useful lines :
./icpSparse -druns a demo with the files stored in the media directory./icpSparse -i1 /path/to/first/objectFile.obj -i2 ../path/to/second/objectFile.obj -o /path/to/output/directory/ -plwould be a minimal line in order to use the point-to-plane variant of the sparse ICP.
Notice that you should always specify if you want to use the point-to-plane variant (-pl) or the point-to-point variant (-po).
There is a python script in the experiments directory in order to reproduce the results presented in the report.
Bouaziz, Sofien, Andrea Tagliasacchi, et Mark Pauly. « Sparse Iterative Closest Point ». In Proceedings of the Eleventh Eurographics/ACMSIGGRAPH Symposium on Geometry Processing, 113–123. SGP ’13. Aire-la-Ville, Switzerland, Switzerland: Eurographics Association, 2013. doi:10.1111/cgf.12178.