Gaussian Process regression
In many practical applications of shape modelling the goal is to infer the full shape from given partial observations. Gaussian Process regression is an inference technique that can be used to predict the unseen part of a shape from such information.
In this video you will learn the mathematics of Gaussian Process regression, how it can be used to predict the most likely shape from an observed part, and how you can obtain a new shape model which is constrained to match the observed part of the shape.
5-3: Gaussian Process regession from NMC Universität Basel on Vimeo.