One of the most important phases of medical diagnostics is the pathology zone extraction from medical images and its further analysis. In relation to this, a task of 3D visualization of the region of interest (biomedical object or a pathology zone) emerges in computer diagnostics. It is important to use the methods of visualization that require a small number of control points, because of the large number of calculations necessary to visualize the object. In this paper a method of contour extraction and selection of control points is proposed. The method provides a minimal number of control points, which are necessary to precisely interpolate the contour. The proposed method has shown stable results even on complex contour forms with boundary concavities.