Segmentation of Abdominal Computed Tomography Scans Using Analysis of Texture Features and Its Application to Personalized Forward Electrocardiography Modeling
Electrical activity of the heart muscles can be measured noninvasively on the body surface using electrocardiography (ECG) method. Numerical results of ECG modeling can be effectively used for diagnosis and treatment planning for cardiovascular diseases. In this paper we propose a method for personalized ECG modeling. One of the most important stages in this process is the generation of personalized torso voxel models, containing heart, lungs, fat, muscles and other organs of the abdominal cavity. Since abdomen segmentation is the most complicated and time consuming task, we propose an automated method for abdominal computed tomography scans (CT) segmentation, based on the texture analysis. We also present numerical analysis of the influence of some anatomical structures of the abdomen on the ECG modeling results.
The research was supported by the Russian Foundation for Basic Research (RFBR) under grants 17-01-00886 and 18-00-01524 (18-00-01661). The authors thank Sechenov University and Institute of Radiology (Rostock, Germany) for providing anonymized CT datasets.
- 3.Hofer, M.: CT Teaching Manual. Georg Thieme Verlag, Stuttgart (2007)Google Scholar