0411420 20240103182326.320060210235959.9 0-7803-9134-9 2 Los Alamitos IEEE 2005 4 s. 578-581 image processing art restoration image registration cav_un_auth*0101125 Kašpar Roman UTIA-B Ústav teorie informace a automatizace AV ČR, v. v. i. cav_un_auth*0216339 Petrů L. CZ cav_un_auth*0101238 Zitová Barbara UTIA-B Department of Image Processing Ústav teorie informace a automatizace AV ČR, v. v. i. cav_un_auth*0101087 Flusser Jan UTIA-B Department of Image Processing Ústav teorie informace a automatizace AV ČR, v. v. i. cav_un_auth*0015832 Hradilová J. CZ 09K GA102/04/0155 GA ČR cav_un_auth*0001806 GA203/04/2091 GA ČR cav_un_auth*0001163 CEZ:AV0Z10750506 Spatial alignment of visible and ultraviolet images of microscopic cross-sections from old artworks is solved by means of mutual information. In the next step, preliminary layer segmentation is applied. In the final part, method for content-based image retrieval is described using cooccurence matrices. The aim is to facilitate search for restorers reports about artworks with similar characteristic, based on corss-section images. Prostorove sesazeni obrazku mikroskopickych rezu starych umeleckych del, porizenych ve viditelnem a ultrafialovem spektru, je reseno pomoci registracni metody "mutual information". V dalsim kroku je vyhotovena prvotni detekce jednotlivych vrstev. Na zaver je popsana metoda pro vyhledavani v databazi na zaklade obrazove informace pomoci kookurencnich matic. Hlavni snahou bylo nabidnout restauratorum snadnejsi vyhledani restauratorskych zprav s podobnou tematikou na zaklade rezu. cav_un_auth*0213246 IEEE International Conference on Image Processing. ICIP'05 /12./ Janov IT 11.09.2005-14.09.2005 JD 2006 ZOI http://hdl.handle.net/11104/0131502 UTIA-B 20050150 2005 2005 Los Alamitos IEEE 0-7803-9134-9 Proceedings of the 12th IEEE International Conference on Image Processing. ICIP'05 578 581