Dynamic boundary location for 2D echocardiographic images in a semi-automated environment

Advanced Algorithmic Approaches to Medical Image Segmentation

“…First, the center of the LV cavity has to be identified before the process begins. In most of the previous articles this is done manually by a human operator using a pointing device [82] [98] . Automatic LV center defining has been addressed by a number of researchers and this will be discussed separately in Section 2.4.…”

“…Different signal processing techniques have been employed to improve the noise sensitivity of edge detection on radial lines by taking advantage of the a priori knowledge of either the imaged geometry or the noise characteristics. For example the anatomical and geometrical information is used for snake initialization in [90] and [98] which is then followed by a dynamic programming-based energy minimization technique to select a particular edge point out of the multiple extracted candidate edges over the radii. In [85] simulated annealing and in [81] and [82] 'improved multi-grid dynamic programming' together with the noise characteristics were used for final edge detection.…”