. . "The aim of this project is to study, analyze and design selected segmentation methods in order to use them when reconstructing the objects from biomedical images. These methods include statistical edge and corner detectors on one hand, and artificial neural networks, on the other hand. The edge detection usually belongs between the first executed steps in image reconstruction. In this way, the primal temporary results, which are further used as an input for more complex and sophisticated segmentation methods, are produced. An example of such complex methods can be mathematical morphology, deformable models or artificial neural networks. The last one is focused on in this project. The statistical edge detectors as well as neural networks can be classified as noise robust methods. Since the biomedical images are full of noise, these two image processing methods are deeply focused on. The crucial part of this project consists in close cooperation with Manchester Metropolitan University (MMU), where"@en . "4"^^ . . "0"^^ . "4"^^ . "Tento projekt m\u00E1 za c\u00EDl studium, anal\u00FDzu a n\u00E1vrh vybran\u00FDch segmenta\u010Dn\u00EDch metod za \u00FA\u010Delem jejich n\u00E1sledn\u00E9ho pou\u017Eit\u00ED p\u0159i rekonstrukci objekt\u016F v biomedic\u00EDnsk\u00FDch obrazech. Tyto metody zahrnuj\u00ED jednak statistick\u00E9 detektory hran a jin\u00FDch v\u00FDznamn\u00FDch jev\u016F v obraze, a jednak um\u011Bl\u00E9 neuronov\u00E9 s\u00EDt\u011B. V algoritmech rekonstrukce obrazu je detekce hran obvykle jedn\u00EDm z prvn\u00EDch nezbytn\u00FDch krok\u016F vedouc\u00EDch k z\u00EDsk\u00E1n\u00ED pr\u016Fb\u011B\u017En\u00FDch v\u00FDsledk\u016F, kter\u00E9 se d\u00E1le pou\u017E\u00EDvaj\u00ED jako vstup sofistikovan\u011Bj\u0161\u00EDch a v\u011Bt\u0161inou na m\u00EDru \u0161it\u00FDch segmenta\u010Dn\u00EDch technik. Mezi tyto techniky m\u016F\u017Eeme za\u0159adit nap\u0159\u00EDklad matematickou morfologii, deformabiln\u00ED modely \u010Di neuronov\u00E9 s\u00EDt\u011B. V tomto projektu se budeme v\u011Bnovat posledn\u00ED ze t\u0159\u00ED v\u00FD\u0161e uveden\u00FDch metod segmentace obrazu, a to um\u011Bl\u00FDm neuronov\u00FDm s\u00EDt\u00EDm. Jak detektory hran zalo\u017Een\u00E9 na statistick\u00FDch testech, tak neuronov\u00E9 s\u00EDt\u011B pat\u0159\u00ED mezi techniky vykazuj\u00EDc\u00ED relativn\u011B malou m\u00EDru citlivosti v\u016F\u010Di \u0161umu, p\u0159\u00EDtomn\u00E9ho v obraze. Proto\u017Ee v biomedic\u00EDnsk\u00FDch obrazech je obvykle dostate\u010Dn\u011B velk\u00E9 mno\u017Estv\u00ED \u0161umu znehodnocuj\u00EDc"@cs . . "Reconstruction of objects from biomedical images using statistic methods and artificial intelligence"@en . . "0"^^ . "Rekonstrukce objekt\u016F v biomedic\u00EDnsk\u00FDch obrazech pomoc\u00ED statistick\u00FDch metod a metod um\u011Bl\u00E9 inteligence"@cs . . . "823"^^ . . "statistick\u00E1 anal\u00FDza, neuronov\u00E9 s\u00EDt\u011B, rekonstrukce objekt\u016F v histologick\u00FDch sn\u00EDmc\u00EDch"@en . . "1"^^ . "563"^^ . .