Aim of this project is the automatic segmentation of 4-dimensional cardiac image data captured through Ultrasound, CT, MRT and other modalities. The research focuses on deformable model methods, which are capable of combining geometry, physics and approximation theory, and are proven more effective in medical applications than the traditional model free methods. Extracting cardiac chambers from the original image data through segmentation is the necessary preprocessing step for the construction of the cardiac anatomical model, which is the starting point for further individually electrophysiological and electromechanical simulation of the heart.