Indexation et recherche d'images par arbres des coupes. (Image indexing and retrieval using component trees)

This thesis explores component trees, hierarchical structures from Mathematical Mor-phology, and their application to image retrieval and related tasks. The distinct com-ponent trees are analyzed and a novel classification into two superclasses is proposed,as well as a contribution to indexing and representation of the hierarchies using den-drograms. The first contribution to the field of image retrieval is in developing a novelfeature detector, built upon the well-established MSER detection. The tree-basedimplementation of the MSER detector allows for changing the underlying tree in or-der to produce features of different stability properties. This resulted in the Tree ofShapes based Maximally Stable Region detector, leading to improvements over MSERin retrieval performance. Focusing on feature description, we extend the concept of2D pattern spectra and adapt their global variant to more powerful, local schemes.Computed on the components of Min/Max-tree, they are histograms holding the in-formation on distribution of image region attributes. The rotation and translationinvariance is preserved from the global descriptor, while special attention is givento achieving scale invariance. We report comparable results to SIFT in image clas-sification, as well as outperforming Morphology-based descriptors in satellite imageretrieval, with a descriptor shorter than SIFT. Finally, a preprocessing or simplifica-tion technique for component trees is also presented, allowing the user to reevaluatethe measures of region level of aggregation imposed on a component tree. The thesisis concluded by outlining the future perspectives based on the content of the thesis.