Co-clustering For Image Category Recognition
This paper presents a novel approach to learning a visual dictionary from sub-spaces, using information theoretic co-clustering, where each sub-space is associated with a semantically relevant part of a visual category. The standard dictionary learning technique, called `Bag-of-Features' is limited by problems of high-dimensionality, sparsity, and noise associated with affine invariant feature descriptors. Our approach draws inspiration from the relation between object part-based models; semantic topic models; non-negative matrix factorization of multivariate data; and sub-spaces in feature space, to resolve these issues in learning a dictionary. We use co-clustering, which performs simultaneous clustering and dimensionality reduction in an optimal way, to discover multiple semantically relevant sub-spaces. Our approach is comprehensively evaluated on several popular datasets: Caltech-101; Caltech-256; Pascal VOC 2006,2007,and 2010; and Scene-15, and is shown to consistently outperform the Bag-of-Features approach.