Liang Mi’s ECCV Paper Got Accepted, Congratulation!

A new clustering method which solves the k-means clustering problem with variational optimal transportation. It leverages power Voronoi diagram to aggregate empirical observations into a fixed number of Voronoi cells while maintaining the minimum transportation cost and preliminary results have shown its applications in domain adaptation, remeshing, and representation learning.

Image of Liang Mi's ECCV Paper: A new clustering method which solves the k-means clustering problem with variational optimal transportation. It leverages power Voronoi diagram to aggregate empirical observations into a fixed number of Voronoi cells while maintaining the minimum transportation cost and preliminary results have shown its applications in domain adaptation, remeshing, and representation learning.