. . . "2015-01-22+01:00"^^ . . "2007-01-01+01:00"^^ . . . . . . "GP102/07/P263" . . "http://www.isvav.cz/projectDetail.do?rowId=GP102/07/P263"^^ . . " MRI images" . "2008-12-31+01:00"^^ . . "2008-04-25+02:00"^^ . "image processing" . . "Neline\u00E1rn\u00ED multimod\u00E1ln\u00ED registrace pro automatickou morfometrii obraz\u016F mozku z MRI zalo\u017Eenou na anatomicky omezen\u00FDch prostorov\u00FDch deformac\u00EDch" . . . "image processing; image registration; MRI images; computational neuroanatomy"@en . "The results of the projects include two methods for nonlinear registration of MRI brain images: 1) block-matching registration 2) registration with smoothing of force fields. The project was solved according to the schedule which was set in the project a"@en . "0"^^ . . . "0"^^ . "2"^^ . . "2"^^ . "1"^^ . "Nonlinear multimodal registration for automatic morphometry of MRI brain images based on anatomically constrained spatial deformations"@en . . "Methods of image registration with the use of nonlinear spatial transformations play a crucial role in the field of computational neuroanatomy, particularly in automatic morphometry of brain images. Voxel-based morphometry includes calculation of local tissue concentrations, whereas direct analyses of spatial transformations computed during the registration are performed in deformation-based morphometry. Limitations of voxel-based morphometry lie in the necessity of tissue classification and gaussian smoothing of binary images, in result of which the spatial resolution of anatomical abnormality detection is unsatisfactory. It is possible to avoid this drawback with the use of precise high-dimensional registration. On the other hand, new anatomic structures may unintentionally arise in the warped image due to the precise matching. Thus, it is necessary to find a method how to prevent this unwanted effect. Additional problems have to be solved, in order to utilize high-dimensional registration in"@en . . "V\u00FDsledkem projektu jsou dv\u011B metody pro neline\u00E1rn\u00ED registraci MRI obraz\u016F mozku: 1) registrace zalo\u017Eena na l\u00EDcov\u00E1n\u00ED podobraz\u016F a 2) registrace s vyhlazov\u00E1n\u00EDm silov\u00FDch pol\u00ED. Pr\u016Fb\u011Bh \u0159e\u0161en\u00ED projektu odpov\u00EDdal zhruba harmonogramu nastaven\u00E9m v projektov\u00E9 p\u0159ihl\u00E1\u0161"@cs . " image registration" . "Metody registrace obraz\u016F s vyu\u017Eit\u00EDm neline\u00E1rn\u00EDch prostorov\u00FDch transformac\u00ED hraj\u00ED z\u00E1sadn\u00ED \u00FAlohu ve v\u00FDpo\u010Detn\u00ED neuroanatomii, zejm\u00E9na p\u0159i automatick\u00E9 morfometrii obraz\u016F mozku, kter\u00E1 je zalo\u017Eena bu\u010F na v\u00FDpo\u010Dtu tzv. koncentrace tk\u00E1n\u011B ve voxelech, nebo na p\u0159\u00EDm\u00E9m hodnocen\u00ED prostorov\u00FDch transformac\u00ED vypo\u010D\u00EDtan\u00FDch b\u011Bhem procesu registrace. Nev\u00FDhody automatick\u00E9 morfometrie zalo\u017Een\u00E9 na voxelech spo\u010D\u00EDvaj\u00ED v nutnosti klasifikovat voxely podle typu mozkov\u00E9 tk\u00E1n\u011B a ve vyhlazov\u00E1n\u00ED bin\u00E1rn\u00EDch obraz\u016F gaussovsk\u00FDm filtrem s n\u00E1slednou ztr\u00E1tou rozli\u0161ovac\u00ED schopnosti p\u0159i prostorov\u00E9 detekci anatomick\u00FDch abnormalit. Pou\u017Eit\u00EDm p\u0159esn\u00E9 mnohorozm\u011Brn\u00E9 registrace je mo\u017En\u00E9 se zm\u00EDn\u011Bn\u00FDm nev\u00FDhod\u00E1m vyhnout. Na druhou stranu v\u0161ak hroz\u00ED, \u017Ee zcela p\u0159esn\u00FDm sl\u00EDcov\u00E1n\u00EDm obraz\u016F m\u016F\u017Ee doj\u00EDtke vzniku nov\u00FDch, v p\u016Fvodn\u00EDm obraze neexistuj\u00EDc\u00EDch struktur. Je proto nutn\u00E9 naj\u00EDt postup, jak tomuto ne\u017E\u00E1douc\u00EDmu jevu zamezit. Dal\u0161\u00ED probl\u00E9my, kter\u00E9 je nutn\u00E9 pro vyu\u017Eit\u00ED mnohorozm\u011Brn\u00E9 registrace vy\u0159e\u0161it, souvisej\u00ED s nalezen\u00EDm vhodn\u00E9ho prostorov\u00E9ho" .