Paper No: 147
Image Similarity based on Hierarchies of ICA Mixtures
Author(s): Arturo Serrano, Addisson Salazar, Jorge Igual, Luis Vergara
This paper presents a novel algorithm to build hierarchies from independent component analyzer mixtures and its application to image similarity measure. The hierarchy algorithm composes an agglomerative (bottom-up) clustering from the estimated parameters (basis vectors and bias terms) of the ICA mixture. Merging at different levels of the hierarchy is made using the Kullback-Leibler distance between clusters. The procedure is applied to merge similar patches on a natural image, to group different images of an object, and to create hierarchical levels of clustering from images of different objects. Re-sults show suitable image hierarchies obtained by clustering from basis func-tions to higher-level structures.