Skip to content

kargarisaac/PointNet-SemSeg-VKITTI3D

Repository files navigation

3D Semantic Segmentation of virtual kitti dataset using PointNet

The main code is from PointNet GitHub Repo

Dataset

You can download the dataset from here.

All files are provided as numpy .npy files. Each file contains a N x F matrix, where N is the number of points in a scene and F is the number of features per point, in this case F=7. The features are XYZRGBL, the 3D XYZ position, the RGB color and the ground truth semantic label L. Each file is for a scene.

Training

Once you have downloaded and prepared data, to start training use main.ipynb.

Visualise data

For data visualization you can use vis_data_vispy.py file.

Selected Projects that Use PointNet

About

VKITTI3D semantic segmentation using PointNet

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published