I wanted to share a project I have been collaborating on with Soumik.
We take the problem of segmenting 3D point clouds that are important for modeling geometric properties from data. Today, we are delighted to open-source our repository that implements the PointNet [1] model family for this purpose. We provide TensorFlow implementations with full support for TPUs and distributed training with mixed-precision (for GPUs). We provide models pre-trained on the four categories of the ShapeNet core dataset [2]. Here’s also a blog post we have prepared to make it easier for getting started.
As always, don’t hesitate to reach out if you have any questions.
References:
[1] [1612.00593] PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation