Isotopic Ventriculography: Detection of the Left Ventricle Using Active Contour Method

There are numerous image segmentation methods. Each has its advantages but is not entirely satisfactory. They must be adapted to the specific medical applications at hand. Active contours are deformable models that eliminate the need to link contour points but require the adjustment of numerous parameters. The active contour we studied is based on gradient diffusion and on parameters of stiffness and elasticity. The method was applied to a database of 128 isotopic ventriculography images to detect the contours of the left ventricle. This approach involved initializing a contour with a sequence of nodes subjected to external forces representing the image characteristics and to internal elasticity and stiffness forces that ensure consistency as the initial contour evolves. This evolution allows the contour to reach a state of equilibrium while minimizing these energies. We obtained good results, with a runtime of approximately 5.15 seconds. This is a fast, accurate, and reliable diagnostic aid.

Pages: 64, Paperback, Our Knowledge Publishing