This paper presents a new approach for 2D object segmentations using an automatic method applied on images with severe noise conditions and locating objects with a very high degree of deformation. We use a physically-based shape model to obtain a deformable template, which is denned on a canonical ortogonal coordinate system. The proposed methodology works from a set of samples and from the output of an edge detector to segment the objects by using a reformulated Hough transform (automatic initialization) together with an optimization procedure (on a learned surface of deformation). Results from biomedical images are presented.
CITATION STYLE
Garrido, A., De La Blanca, N. P., & Garc Ί A-Silvente, M. (1997). A new methodology to automatically segment biomedical images. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1311, pp. 372–379). Springer Verlag. https://doi.org/10.1007/3-540-63508-4_145
Mendeley helps you to discover research relevant for your work.