"[4C986DBCFDC8]" . "Proceedings of 10th International Conference Process Control 2012" . "Pardubice" . "Segmentation of Overlapping Biomedical Objects" . "overlapping objects; region growing method; image segmentation; digital filters; morphology operation; biomedical image processing"@en . . "Segmentation of Overlapping Biomedical Objects" . . . "2012-06-11+02:00"^^ . "Segmentation of Overlapping Biomedical Objects"@en . "RIV/60461373:22340/12:43893348!RIV13-MSM-22340___" . "12"^^ . . "Z(MSM6046137306)" . "Yadollahi, Mohammadreza" . . "The paper presents an algorithm for identifying common boundary between two connected neighbouring regions in the orthodontic image. The digital image is provided by top source illumination on plaster cast of the orthodontic image. The initial stage of image processing is based on segmentation of the object from its background by Otsu method. This segmentation provides a complete background and makes it possible to focus only on the object for further image processing. In the secondary stage, de-nosing and intensity adjustment are applied on plaster cast image as an enhancement method. In the third stage, the region growing segmentation is performed by selecting the appropriate seed point and identifying the similarity criteria for segmentation of every region in the object. The last stage includes the algorithm presented here which is complementary to the region growing segmentation for identifying the common boundary between two connected neighbouring regions. The proposed algorithm is performed by extracting the boundary, identifying the convex deficiency of a shape and applying the mathematical methods on two connected neighbourhood regions. The main goal of this paper is (a) to present the mathematical methods for image analysis in the orthodontic treatment, (b) to present the use of region growing method for image segmentation and (c) to propose the algorithm for identifying common boundary of two neighbouring regions.the orthodontic treatment, (b) to present the use of region growing method for image segmentation and (c) to propose the algorithm for identifying common boundary of two neighbouring regions."@en . "Segmentation of Overlapping Biomedical Objects"@en . "Kouty nad Desnou" . . "The paper presents an algorithm for identifying common boundary between two connected neighbouring regions in the orthodontic image. The digital image is provided by top source illumination on plaster cast of the orthodontic image. The initial stage of image processing is based on segmentation of the object from its background by Otsu method. This segmentation provides a complete background and makes it possible to focus only on the object for further image processing. In the secondary stage, de-nosing and intensity adjustment are applied on plaster cast image as an enhancement method. In the third stage, the region growing segmentation is performed by selecting the appropriate seed point and identifying the similarity criteria for segmentation of every region in the object. The last stage includes the algorithm presented here which is complementary to the region growing segmentation for identifying the common boundary between two connected neighbouring regions. The proposed algorithm is performed by extracting the boundary, identifying the convex deficiency of a shape and applying the mathematical methods on two connected neighbourhood regions. The main goal of this paper is (a) to present the mathematical methods for image analysis in the orthodontic treatment, (b) to present the use of region growing method for image segmentation and (c) to propose the algorithm for identifying common boundary of two neighbouring regions.the orthodontic treatment, (b) to present the use of region growing method for image segmentation and (c) to propose the algorithm for identifying common boundary of two neighbouring regions." . . . "Yadollahi, Mohammadreza" . . "Proch\u00E1zka, Ale\u0161" . "RIV/60461373:22340/12:43893348" . . "978-80-7395-500-7" . . "167233" . "2"^^ . . . "Univerzita Pardubice" . . "2"^^ . "22340" . . . . . . .