Discrete Inference and Learning in Artificial Vision
This article will be about the probability graphical models. At first, I will just gain resources link:
- A survey paper:
Markov Random Field modeling, inference & learning in computer vision
& image understanding: A survey
2. Nikos Paragios, Professor of Applied Mathematics & Computer Science

where the energy (or cost, objective) function E(x, D; w) can be regarded as a quality measure of a parameter configuration x in the solution space given the observed data D, and w denotes the model parameters
visual perception involves three main tasks: modeling, inference and learning. The modeling has to accomplish: (i) the choice of an appropriate representation of the solution using a tuple of variables x; and (ii) the design of the class of energy functions E(x, D; w) which can correctly mea- sure the connection between x and D. The inference has to search for the configuration of x leading to the optimum of the energy function, which corresponds to the solution of the original problem. The learning aims to select the optimal model parameters w based on the training data.
3. Conditional Random Fields
Fully-connected CRFs: this includes codes in python. This could be a good start to combine CRF to TensorFlow

