PointCloud value prediction - regression

I have point clouds for 400 objects. I have also some other features of these objects. For example their weights.

How I can now train model to predict the weight based on the shape(point cloud)?

I have seen pointnet but is seems it is only for classification?

Topic 3d-reconstruction cloud data

Category Data Science


My understanding of your problem s the target is weight and the point cloud are the features.

Creating a different representations of the point cloud that are amenable to machine learning is called feature engineering.

Possible feature engineering options:

  • Input each individual cloud point.
  • Calculate total volume of cloud points.
  • Template matching. Define shapes and then categorize each point cloud into one of those shapes. The shapes could be for the entire point cloud or subsets.

About

Geeks Mental is a community that publishes articles and tutorials about Web, Android, Data Science, new techniques and Linux security.