Text this: Sparse Representation, Modeling and Learning in Visual Recognition