In a large number of applications, the processing relies on objects or area of interests, and the pixel-based image representation is notwell adapted. These applications would benefit from a region-based processing. Early examples of region-based processing can be found in the area of image segmentation, such as the quad tree. Recently, in mathematical morphology, the connected operators have received much attention. They are region-based filtering tools that act by merging flat zones. They have good contour preservation properties in the sense that they do not create any new boundaries, neither do they shift the existing ones. One popular implementation for connected operators relies on tree-based image representations, notably threshold decomposition representations and hierarchical representations. Those tree-based image representations are widely used in many image processing and computer vision applications. Tree-based connected operators consist in constructing a set of nested or disjoint connected components, followed by a filtering of these connected components based on an attribute function characterizing each connected component. Finally, the filtered image is reconstructed from the simplified tree composed of the remaining connected components. In the work presented in this thesis, we propose to expand the ideas of tree-based connected operators. We introduce the notion of tree-based shape spaces, built from tree-based image representations. Many state-of-the-art methods relying on tree-based image representations consist of analyzing this shape space. A first consequence of this change of point of view is our proposition of a local feature detector, called the tree-based Morse regions (TBMR). It can be seen as a variant of the MSER method. The selection of TBMRs is based on topological information, and hence it extracts the regions independently of the contrast, which makes it truly contrast invariant and quasi parameters free. The accuracy and robustness of the TBMR approach are demonstrated by the repeatability test and by applications to image registration and 3D reconstruction, as compared to some state-of-the-art methods. The basic idea of the main proposition in this thesis is to apply connected filters on the shape space. Such a processing is called the framework of shape-based morphology. It is a versatile framework that deals with region-based image representations. It has three main consequences. 1) For filtering purpose, it is a generalization of the existing tree-based connected operators. Indeed, the framework encompasses classical existing connected operators by attributes. Besides, It also allows us to propose two classes of novel connected operators: shape-based lower/upper levelings and shapings. 2) This framework can be used to object detection/segmentation by selecting relevant points in the shape space. 3) We can also use this framework to transform the hierarchies using the extinction values, so that a hierarchical simplification or segmentation is obtained. Some applications are developed using the framework of shape-based morphology to demonstrate its usefulness. The applications of the shape-based filtering to retinal image analysis show that a mere filtering step that we compare to more evolved processings, achieves state-of-the-art results. An efficient shaping used for image simplification is proposed by minimizing Mumford-Shah functional subordinated to the topographic map. For object detection/segmentation, we proposed a context-based energy estimator that is suitable to characterize object meaningfulness. Last, we extend the hierarchy of constrained connectivity using the aspect of hierarchy transformation of constrained connectivity using the aspect ofhierarchy transformation.

Dans le travail présenté dans cette thèse, nous proposons d'élargir les idées des opérateurs connexes à base d'arbres. Nous introduisons la notion d'espaces de formes à base d'arbres, construit à partir des représentations d'image à base d'arbres. De nombreuses méthodes de l'état de l'art, s'appuyant sur ces représentations d'images à base d'arbres, consistent à analyser cet espace de forme. Une première conséquence de ce changement de point de vue est notre proposition d'un détecteur de caractéristiques locales, appelé les « tree-based Morse regions » (TBMR). Cette approche peut être considérée comme une variante de la méthode des MSER. La sélection des TBMRs est basée sur des informations topologiques, et donc extrait les régions indépendamment du contraste, ce qui la rend vraiment invariante aux changements de contraste; de plus, la méthode peut être considérée sans paramètres. La précision et la robustesse de l'approche TBMR sont démontrées par le test de reproductibilité et par des applications au recalage d'image et à la reconstruction 3D, en comparaison des méthodes de l'état de l'art. L'idée de base de la proposition principale dans cette thèse est d'appliquer les opérateurs connexes à l'espace des formes. Un tel traitement est appelé la morphologie basée sur la forme. Ce cadre polyvalent traite des représentations d'images à base de région. Il a trois conséquences principales. 1) Dans un but de filtrage, il s'agit d'une généralisation des opérateurs connexes à base d'arbres. En effet, le cadre englobe les opérateurs connexes classiques par attributs. En outre, il permet également de proposer deux nouvelles classes d'opérateurs connexes: nivellements inférieurs/supérieurs à base de forme et shapings. 2) Ce cadre peut être utilisé pour la détection/segmentation d'objets en sélectionnant les points pertinents dans l'espace des formes. 3) Nous pouvons également utiliser ce cadre pour transformer les hiérarchies en utilisant les valeurs d'extinction, obtenant ainsi une simplification/segmentation hiérarchique. Afin de montrer l'utilité de l'approche proposée, plusieurs applications sont développées. Les applications à l'analyse d'images rétinenne de filtrage basé sur la forme montrent qu'une simple étape de filtrage, comparée à des traitements plus évolués, réalise des résultats au niveau de l'état de l'art. Une application de shaping pour la simplification d'image est proposée, fondée sur une minimisation de la fonctionnelle de Mumford-Shah subordonnée à l'arbre de formes. Pour la détection/segmentation d'objets, nous proposons un estimateur de l'énergie basée sur le contexte. Cet estimateur est approprié pour caractériser la signification d'objet. Enfin, nous étendons le cadre de la connectivité contrainte en utilisant l'aspect de transformation de hiérarchie / In a large number of applications, the processing relies on objects or area of interests, and the pixel-based image representation is notwell adapted. These applications would benefit from a region-based processing. Early examples of region-based processing can be found in the area of image segmentation, such as the quad tree. Recently, in mathematical morphology, the connected operators have received much attention. They are region-based filtering tools that act by merging flat zones. They have good contour preservation properties in the sense that they do not create any new boundaries, neither do they shift the existing ones. One popular implementation for connected operators relies on tree-based image representations, notably threshold decomposition representations and hierarchical representations. Those tree-based image representations are widely used in many image processing and computer vision applications. Tree-based connected operators consist in constructing a set of nested or disjoint connected components, followed by a filtering of these connected components based on an attribute function characterizing each connected component. Finally, the filtered image is reconstructed from the simplified tree composed of the remaining connected components. In the work presented in this thesis, we propose to expand the ideas of tree-based connected operators. We introduce the notion of tree-based shape spaces, built from tree-based image representations. Many state-of-the-art methods relying on tree-based image representations consist of analyzing this shape space. A first consequence of this change of point of view is our proposition of a local feature detector, called the tree-based Morse regions (TBMR). It can be seen as a variant of the MSER method. The selection of TBMRs is based on topological information, and hence it extracts the regions independently of the contrast, which makes it truly contrast invariant and quasi parameters free. The accuracy and robustness of the TBMR approach are demonstrated by the repeatability test and by applications to image registration and 3D reconstruction, as compared to some state-of-the-art methods. The basic idea of the main proposition in this thesis is to apply connected filters on the shape space. Such a processing is called the framework of shape-based morphology. It is a versatile framework that deals with region-based image representations. It has three main consequences. 1) For filtering purpose, it is a generalization of the existing tree-based connected operators. Indeed, the framework encompasses classical existing connected operators by attributes. Besides, It also allows us to propose two classes of novel connected operators: shape-based lower/upper levelings and shapings. 2) This framework can be used to object detection/segmentation by selecting relevant points in the shape space. 3) We can also use this framework to transform the hierarchies using the extinction values, so that a hierarchical simplification or segmentation is obtained. Some applications are developed using the framework of shape-based morphology to demonstrate its usefulness. The applications of the shape-based filtering to retinal image analysis show that a mere filtering step that we compare to more evolved processings, achieves state-of-the-art results. An efficient shaping used for image simplification is proposed by minimizing Mumford-Shah functional subordinated to the topographic map. For object detection/segmentation, we proposed a context-based energy estimator that is suitable to characterize object meaningfulness. Last, we extend the hierarchy of constrained connectivity using the aspect of hierarchy transformation of constrained connectivity using the aspect ofhierarchy transformation.

In a large number of applications, the processing relies on objects or area of interests, and the pixel-based image representation is notwell adapted. These applications would benefit from a region-based processing. Early examples of region-based processing can be found in the area of image segmentation, such as the quad tree. Recently, in mathematical morphology, the connected operators have received much attention. They are region-based filtering tools that act by merging flat zones. They have good contour preservation properties in the sense that they do not create any new boundaries, neither do they shift the existing ones. One popular implementation for connected operators relies on tree-based image representations, notably threshold decomposition representations and hierarchical representations. Those tree-based image representations are widely used in many image processing and computer vision applications. Tree-based connected operators consist in constructing a set of nested or disjoint connected components, followed by a filtering of these connected components based on an attribute function characterizing each connected component. Finally, the filtered image is reconstructed from the simplified tree composed of the remaining connected components. In the work presented in this thesis, we propose to expand the ideas of tree-based connected operators. We introduce the notion of tree-based shape spaces, built from tree-based image representations. Many state-of-the-art methods relying on tree-based image representations consist of analyzing this shape space. A first consequence of this change of point of view is our proposition of a local feature detector, called the tree-based Morse regions (TBMR). It can be seen as a variant of the MSER method. The selection of TBMRs is based on topological information, and hence it extracts the regions independently of the contrast, which makes it truly contrast invariant and quasi parameters free. The accuracy and robustness of the TBMR approach are demonstrated by the repeatability test and by applications to image registration and 3D reconstruction, as compared to some state-of-the-art methods. The basic idea of the main proposition in this thesis is to apply connected filters on the shape space. Such a processing is called the framework of shape-based morphology. It is a versatile framework that deals with region-based image representations. It has three main consequences. 1) For filtering purpose, it is a generalization of the existing tree-based connected operators. Indeed, the framework encompasses classical existing connected operators by attributes. Besides, It also allows us to propose two classes of novel connected operators: shape-based lower/upper levelings and shapings. 2) This framework can be used to object detection/segmentation by selecting relevant points in the shape space. 3) We can also use this framework to transform the hierarchies using the extinction values, so that a hierarchical simplification or segmentation is obtained. Some applications are developed using the framework of shape-based morphology to demonstrate its usefulness. The applications of the shape-based filtering to retinal image analysis show that a mere filtering step that we compare to more evolved processings, achieves state-of-the-art results. An efficient shaping used for image simplification is proposed by minimizing Mumford-Shah functional subordinated to the topographic map. For object detection/segmentation, we proposed a context-based energy estimator that is suitable to characterize object meaningfulness. Last, we extend the hierarchy of constrained connectivity using the aspect of hierarchy transformation of constrained connectivity using the aspect ofhierarchy transformation.