An Evaluation of a Combined Spline Based and Rate of Energy Dissipation Based Segmentation of Medical Blood Flow Images

Domain characteristics provide some constraints for accurate segmentation of special domain (e.g. flow) images. These constraints usually obtain a type of region based segmentation and in combination with the intensity based segmentation methods obtain overcoming results. When a boundary based segmentation method is considering to be aligned with these region based methods, some special efforts are necessary to make them compatible. Here we address a method to combine two boundary and regional based segmentation methods. In this regard boundary based segmentation is changed to be compatible with a regional based energy dissipation constraint in a sample flow domain. A quadratic spline function is used for boundary based segmentation of flow domain. The regional equivalent factor of this method is provided using the Gauss theorem. The numerical implementation of this method is provided with construction of boundary elements on the spline boundary segments. The rate of energy dissipation at blood elements is applied as the regional segmentation factor. It calculated from the velocity data of the flow Image. An overcoming segmentation factor is provided by mixing both methods. For implementation of this method a simulated Phase- Contrast Magnetic Resonance Image (PC-MRI) of Couette flow system is considered.