3D Point Cloud classification using Deep Learning
Motivation and scope
Point clouds are a set of points in space that represents objects in an environment. They can be acquired directly from 3D scanners or with 3D reconstruction using multiple 2D images. Semantic segmentation of point clouds is the task of classifying each point into object categories. Recent scientific progress in deep learning has resulted in a number of powerful machine learning algorithms for semantic segmentation for both 2D images and 3D point clouds.
The task is to use recent state-of-the-art algorithms for classifying each point in a point cloud. The point clouds can either be of an indoor environment taken with LIDAR scanners or of an outdoor environment from a 3D reconstruction from images taken with a drone.
Good programming skills in at least one of the following: Python, Tensorflow, Keras, PyTorch, and/or Matlab. Basic knowledge in machine learning.
You will learn how to work on a challenging machine learning project using deep learning algorithms for point cloud classification.