Robust Recognition via Information Theoretic Learning

this work brings them together, attempts to impart the theory, optimization and usage of information entropy.

The authors resort to a new information theoretic concept, correntropy, as a robust measure and apply it to solve robust face recognition and object recognition problems.