05376163

a

05376163 Demi, Marcello Bianchini, Elisabetta Faita, Francesco Gemignani, Vincenzo Contour tracking when two gray-level discontinuities are close to each other. Ruiz-Shulcloper, Jos\'e (ed.) et al., Progress in pattern recognition, image analysis and applications. 13th Iberoamerican congress on pattern recognition, CIARP 2008, Havana, Cuba, September 9--12, 2008. Proceedings. Berlin: Springer (ISBN 978-3-540-85919-2/pbk). Lecture Notes in Computer Science 5197, 585-592 (2008). 2008 Berlin: Springer EN contour tracking vascular images ultrasound imaging

doi:10.1007/978-3-540-85920-8_71

Summary: Vascular measurements are indispensable to quantify important indexes of cardiovascular risk and image processing systems are needed to automatically track the vascular structures through sequences of echographic images. Given a starting contour $c _{n }$ on frame $f _{n }$, a contour tracking algorithm is generally based on the application of a mathematical operator at the points of $c _{n }$ and on an iterative procedure which brings such points to the respective points of the contour $c _{n + 1}$ on the subsequent frame $f _{n + 1}$. In this paper, the performances of a mathematical operator which looks for similar regional gray level distributions are compared to those of an edge detection operator. The paper shows that when two or more gray-level discontinuities are present and close to each other, as in the case of arteries, both operators should be used sequentially.