Variational Inference: Ising Model – Towards Data Science

This article focuses on Variational Inference (VI) for the Ising model in application to binary image denoising. As an example, consider a noisy gray scale image on the left and the denoised binary image on the right in the figure above. The Ising model is an example of a Markov Random Field (MRF) and it originated from statistical physics. The Ising model assumes that we have a grid of nodes, where each node can be in one of two states. In the case of binary images, you can think of each node as being a pixel with a black or white color.