3DSmoothNet

Generate the Smoothed Density Valie (SVD) voxel grid and keypoints (saved as .pcd) using the voxelize_all.py. Modify the input and output direcotries at the beginning of hte file. The parameters for the SVD grid are stored in the voxel_parameters.yaml file. Do not modify the directory parameters in the .yaml file, they are update autmatically by the script.
Use the run_benchmark_global.py to generate the features from the SVD rapresentation.
Use the join_all.py script to merge keypoints and features in a single .npz file.

SLAM example

voxelize_tum_slam.py
run_benchmark_global.py (modify the )
join_feature_keypoints.py --cnn_dim 64 -f /benchmark/tum_slam_comparison/3DSmoothNet/features/pioneer_slam2_025/ -o /benchmark/tum_slam_comparison/3DSmoothNet/features/pioneer_slam2_025/

Name		Name	Last commit message	Last commit date
Latest commit History 32 Commits
core		core
data		data
evaluation		evaluation
figures		figures
logs		logs
models		models
.gitignore		.gitignore
3DSmoothNet		3DSmoothNet
CMakeLists.txt		CMakeLists.txt
LICENSE		LICENSE
README.md		README.md
README_old.md		README_old.md
benchmark_helper.py		benchmark_helper.py
demo.py		demo.py
exec_on_benchmark.py		exec_on_benchmark.py
install_pcl.sh		install_pcl.sh
join_all.py		join_all.py
join_feature_keypoints.py		join_feature_keypoints.py
main.cpp		main.cpp
main_cnn.py		main_cnn.py
requirements.txt		requirements.txt
run_benchmark_global.py		run_benchmark_global.py
voxel_parameters.yaml		voxel_parameters.yaml
voxelize_all.py		voxelize_all.py
voxelize_tum_slam.py		voxelize_tum_slam.py