<> "The repository administrator has not yet configured an RDF license."^^ . <> . . "Morphological Analysis for Object Recognition, Matching, and Applications"^^ . "This thesis deals with the detection and classifcation of objects in visual images and with\r\nthe analysis of shape changes between object instances. Whereas the task of object recognition\r\nfocuses on learning models which describe common properties between instances of\r\na specific category, the analysis of the specific differences between instances is also relevant\r\nto understand the objects and the categories themselves. This research is governed by the\r\nidea that important properties for the automatic perception and understanding of objects\r\nare transmitted through their geometry or shape. Therefore, models for object recognition\r\nand shape matching are devised which exploit the geometry and properties of the objects,\r\nusing as little user supervision as possible.\r\nIn order to learn object models for detection in a reliable manner, suitable object representations\r\nare required. The key idea in this work is to use a richer representation of the\r\nobject shape within the object model in order to increase the description power and thus\r\nthe performance of the whole system. For this purpose, we first investigate the integration\r\nof curvature information of shapes in the object model which is learned. Since natural\r\nobjects intrinsically exhibit curved boundaries, an object is better described if this shape\r\ncue is integrated. This subject extends the widely used object representation based on\r\ngradient orientation histograms by incorporating a robust histogram-based description of\r\ncurvature. We show that integrating this information substantially improves detection\r\nresults over descriptors that solely rely upon histograms of orientated gradients.\r\nThe impact of using richer shape representations for object recognition is further investigated\r\nthrough a novel method which goes beyond traditional bounding-box representations\r\nfor objects. Visual recognition requires learning object models from training data. Commonly,\r\ntraining samples are annotated by marking only the bounding-box of objects since\r\nthis appears to be the best trade-off between labeling information and effectiveness. However,\r\nobjects are typically not box-shaped. Thus, the usual parametrization of objects\r\nusing a bounding box seems inappropriate since such a box contains a significant amount\r\nof background clutter. Therefore, the presented approach learns object models for detection\r\nwhile simultaneously learning to segregate objects from clutter and extracting their\r\noverall shape, without however, requiring manual segmentation of the training samples.\r\nShape equivalence is another interesting property related to shape. It refers to the ability\r\nof perceiving two distinct objects as having the same or similar shape. This thesis\r\nalso explores the usage of this ability to detect objects in unsupervised scenarios, that is\r\nwhere no annotation of training data is available for learning a statistical model. For this\r\npurpose, a dataset of historical Chinese cartoons drawn during the Cultural Revolution\r\nand immediately thereafter is analyzed. Relevant objects in this dataset are emphasized\r\nthrough annuli of light rays. The idea of our method is to consider the different annuli as\r\nshape equivalent objects, that is, as objects sharing the same shape and devise a method\r\nto detect them. Thereafter, it is possible to indirectly infer the position, size and scale of\r\nthe emphasized objects using the annuli detections.\r\nNot only commonalities among objects, but also the specific differences between them are\r\nperceived by a visual system. These differences can be understood through the analysis\r\nof how objects and their shape change. For this reason, this thesis also develops a novel\r\nmethodology for analyzing the shape deformation between a single pair of images under\r\nmissing correspondences. The key observation is that objects cannot deform arbitrarily,\r\nbut rather the deformation itself follows the geometry and constraints imposed by the\r\nobject itself. We describe the overall complex object deformation using a piecewise linear\r\nmodel. Thereby, we are able to identify each of the parts in the shape which share the same deformation. Thus, we are able to understand how an object and its parts were\r\ntransformed. A remarkable property of the algorithm is the ability to automatically estimate\r\nthe model complexity according to the overall complexity of the shape deformation.\r\nSpecifically, the introduced methodology is used to analyze the deformation between original\r\ninstances and reproductions of artworks. The nature of the analyzed alterations ranges\r\nfrom deliberate modifications by the artist to geometrical errors accumulated during the\r\nreproduction process of the image. The usage of this method within this application shows\r\nhow productive the interaction between computer vision and the field of the humanities is.\r\nThe goal is not to supplant human expertise, but to enhance and deepen connoisseurship\r\nabout a given problem."^^ . "2013" . . . . . . . "Juan Antonio"^^ . "Monroy Kuhn"^^ . "Juan Antonio Monroy Kuhn"^^ . . . . . . "Morphological Analysis for Object Recognition, Matching, and Applications (PDF)"^^ . . . "thesis.pdf"^^ . . . "Morphological Analysis for Object Recognition, Matching, and Applications (Other)"^^ . . . . . . "indexcodes.txt"^^ . . . "Morphological Analysis for Object Recognition, Matching, and Applications (Other)"^^ . . . . . . "lightbox.jpg"^^ . . . "Morphological Analysis for Object Recognition, Matching, and Applications (Other)"^^ . . . . . . "preview.jpg"^^ . . . "Morphological Analysis for Object Recognition, Matching, and Applications (Other)"^^ . . . . . . "medium.jpg"^^ . . . "Morphological Analysis for Object Recognition, Matching, and Applications (Other)"^^ . . . . . . "small.jpg"^^ . . "HTML Summary of #15663 \n\nMorphological Analysis for Object Recognition, Matching, and Applications\n\n" . "text/html" . . . "004 Informatik"@de . "004 Data processing Computer science"@en . .