Optical character recognition using transfer learning decision forests

In this paper, we present a novel method for transfer learning which uses decision forests, and we apply it to recognize characters. We introduce two extensions into the decision forest framework in order to transfer knowledge from the source tasks to a given target task. We show that both of them are important to achieve higher recognition rates. Our experiments demonstrate improvements over traditional decision forests in the MNIST dataset. They also compare favorably against other state-of-the-art classifiers.