While deep neural networks have tackled image classification they lack the interpretability  and it is difficult to guarantee that the convergence, stability. From another perspective, purely geometric algorithms for instance nearest neighbors may not be optimal albeit they are stable and converge under suitable assumptions. Classical machine learning algorithms combined with representation generated by the neural networks can give a reasonable performance. Similar to this one, Voronoi cell can be combined with Bayesian 

Introduction and Related Work (Voronoi Boundary Classification: A High-Dimensional Geometric Approach via Weighted Monte Carlo Integration)

The research paper "Voronoi Boundary Classification: A High-Dimensional Geometric Approach via Weighted Monte Carlo Integration" is structured into four primary sections: (1) Introduction and the Related Work, (2) Methodology and the Algorithm, (3) Experimental Evaluation, and (4) Conclusions and Future Work.

San Diego State University

Google

1. Voronoi Boundary Classification: A High-Dimensional Geometric Approach
via Weighted Monte Carlo Integration

Research Papers for Voronoi Boundary Classification: A High-Dimensional Geometric Approach via Weighted Monte Carlo Integration

http://proceedings.mlr.press/v97/polianskii19a/polianskii19a.pdf

Learn Before

Related

Learn After