The Key Algorithm Implementation And System Development Of The Virtual Angioscopy

Viewing the inner part of organs is very important in clinic, and now many technologies have been developed to achieve it. Endoscopy is an important one. Endoscopy technology has a physical instrument with camera inserted in to the organ and gets images of the inner organ. Obviously, traditional endoscopy hurts patients and it can only get the inner surface image. Recently, with the rapid progress of computer graphics and medical image technologies, Virtual Endoscopy(VE) technology have been proposed and improved a lot since then.Virtual Angioscopy (VA) is a kind of VE which focus on human blood vessel based on the Computed Tomography(CT) or Magnetic Resonance Image(MRI) images. By now, very few researches have been done in VA, but some systematic applications have come to clinic use.Because of the large size of medical image data and the complexity of image computing, early endoscopy systems were implemented mainly on professional graph stations. With the progress of PC hardware, PC based VA system research will cut down system cost significantly and enhance its clinic value a lot. This paper proposes a PC based interactive virtual Angioscopy system and it's key algorithms. In this paper, First, it gives a brief introduction to the history and improvement of virtual Endoscopy system. Then, it presents the process and architecture of virtual angioscopy system. Finally, it discusses the main functions and researches in this paper: (1) In data segmentation part, it introduces the popular 3D medical image segmentation methods and proposes a fast segmentation algorithm for blood vessel data.(2) In the virtual navigation function, it proposes an efficient centerline extraction algorithm that is base on 3D 6-subiteration curve thinning algorithm, with the centerline, our VA system provides an interactive navigation. Also, we provide a centerline extraction algorithm based on the blood vessel's surface mesh.(3) In the rendering part, this paper introduces the marching cubes algorithm that reconstructs the blood vessel inside surface model for rendering at first, discusses some volume rendering methods, and then proposes an improved volume rendering algorithm using GPU acceleration.