Cardiovascular Computer-Aided Diagnosis System Research Based On Visualization Technology

Due to the complexity of cardiovascular anatomy, it makes cardiovascular diseases(CAD) suffer from laborious issue. Though, volumetric visualization of cardiovascular anatomy is achieved by existing computer aided diagnosis system. However, many urgent problems starve for solutions in practical applications, such as effective transfer function designing, high quality and speed visualization rendering and friendly interaction mode.To address above drawbacks, a n improved computer aided CAD diagnosis system is designed using GPU-based volume visualization techniques. The main contributions of this paper are summarized as follows: 1) In order to implement high-quality volumetric data visualization, ray casting method is employed. The method has some superiority such as being similar to natural retinal imaging, high quality rendering and high speed bas ed on GPU parallel capability. Moreover, the illumination and optimization algorithm is embedded in the implementation process, for instance early termination of light and space jump. 2) For displaying the coronary artery, a novel transfer function based on semi-automatic pre-segmentation is proposed. With vascular specified improved region growing method, initial contour is obtained, which is converted to zero level set for iterative calculation. Finally, color assignment is accomplished by weight transfer function. 3) In this paper, the ASM model is employed to segment the outer membrane of the left ventricle. After expert sample collection, the average model is derived, which is used to retrieve the desired segmentation result. 4) Through user interaction mode, regular and irregular cutting is applied for volumetric data. And improved ray casting method is used for rendering. Furthermore, 3D heart mesh generation is executed automatically. 5) The computer aided diagnosis system integrates above technologies to solve the current challenges. The paper introduces the system framework and each functional module. Through performance comparison, the system can provide elegant result and diagnostic reference to clinicians.