npm3d-challenge

MiniChallenge in 2020 NPM3D course

The data layout is:

data/MiniChallenge
- training
- test

The 0 label corresponds to Unclassified data points in the point clouds.

Score

The prevous' year best score is 63.32

We use this implementation of PointNet++ in PyTorch: https://github.com/erikwijmans/Pointnet2_PyTorch

We use "". Grab the code:

wget http://mouse.cs.uwaterloo.ca/code/gco-v3.0.zip

We use CMake as a build system for the C++ code. Build the code:

mkdir build/
cd build/
cmake ../gco
make

You can then check that the example file compiles:

./build/Main

Blomley et al. SHAPE DISTRIBUTION FEATURES FOR POINT CLOUD ANALYSIS- A GEOMETRIC HISTOGRAM APPROACH ON MULTIPLE SCALES https://www.isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/II-3/9/2014/isprsannals-II-3-9-2014.pdf
Loic Landrieu, Hugo Raguet, Bruno Vallet, Clément Mallet, Martin Weinmann. A structured regularization framework for spatially smoothing semantic labelings of 3D point clouds. ISPRS Journal of Photogrammetry and Remote Sensing, Elsevier, 2017, 132, pp.102-118. Link
Bergstra, J., Yamins, D., Cox, D. D. (2013) Making a Science of Model Search: Hyperparameter Optimization in Hundreds of Dimensions for Vision Architectures. To appear in Proc. of the 30th International Conference on Machine Learning (ICML 2013).