Topological Data Analysis: potential applications to computer vision

Topological Data Analysis quantifies hidden topological structures in big raw noisy data. The flagship tool (persistent homology) summarises the underlying structure across all scales. The stability result says that such a summary (persistence diagram) is robust to noise. We shall review potential approaches to combine topological data analysis and machine learning for practical problems in computer vision.