UN ENFOQUE INTEGRADO DE SEGMENTACIÓN PARA EXTRACCIÓN DE SUPERFICIES EN IMÁGENES DIGITALES 3D

M. DEL FRESNO; M. VÉNERE; A. CLAUSSE

La Rioja

Congreso; XIV Congreso Argentino de Ciencias de la Computación; 2008

CACIC

This paper describes an efficient novel strategy for the segmentation of surfaces contained in 3D images, based on deformable models or snakes. The proposed method starts by determining a rough but robust approximation of the objects of  interest using the results of a region-growing algorithm. Then, the closed surface mesh that encloses the region  is constructed and used as the initial geometry of a deformable model for the final refinement. This integrated strategy provides an alternative solution to  one of the flaws of traditional  deformable models, achieving good refinements of internal surfaces in few steps and times much reasonable than the reported by others authors. Experimental segmentation results of complex structures on both simulated and real data are presented, and the method is assessed by comparing with standard reference segmentations. The obtained results show that the approach achieves high quality segmentations with low computational complexity.