kNN approaches in geometric algorithms

Abstract

The idea of kNN (finding the k-th nearest neighbor) is used in many scientific branches, primarily in machine learning, recognition and mathematical statistics. We would like to show further applications in geometric algorithms dealing with point clouds or massive sample data.

We will show results from visualization and shape reconstruction (geometric inference).

Keywords

• kNN
• K nearest neighbors
• nonparametric density estimation
• classification
• machine learning
• supervised learning
• geometric inference