3D Medical image processing

Overview of 3D Medical image processing and visualization

Medical image, such as CT, MRI, shows the information inside the patient body by non-invasive method, so that it is much helpful for doctor´s diagnoses and less painful for patients. However the raw data can only give the material to doctor, the doctor as to decide by himself which is important which is not. The computer-aided diagnoses is to use computer to process the medical to extract the useful information so that the doctor can make a diagnoses decision easier and quicker. In our project the team of Prof. Dr. Christoph Meinel want to do medical image segmentation in three dimension and visualize the segmented object, further more give doctor the means of measurement and analysis of interested object.

Project Description

  • TeleMed-VS   
    • Introduction : TeleMed-VS is one medical image visualization and segmentation tool. It is like an eye to aid doctors in medical image analysis and in diagnosis, who use medical images in neurology, radiology, surgery, or any other field. For image visualization, it renders 3D image data both in slice mode and volume mode. In slice mode, TeleMed-VS displays image in three slice windows, whose are views in the orthogonal axial, coronal and sagittal planes. In volume mode, the tool shows image data as surface shell and semi-translucent object. For image segmentation, it realizes semi-automatic image segmentation mehtod that it only needs a few manual work to get the segmentation result. The preciseness of segmentation is satisfied for medical use.
    • Algorithm: For medical image processing system, finding out the area that interested by physicians Image segmentation is primary work for anatomical analysis and pathological diagnosis. However, it is still challenge and unsolved problem. In this thesis we present a new combined approach designed for automated segmentation of radiological image, such as CT, MRI, etc, to get the organ or interested area from the image. This approach integrates region-based method and boundary-based method. Such integration reduces the drawbacks of both methods and enlarges the advantages of them.  
    • Screenshot:




  • See Through
    • Introduction : Medical image segmentation is the basic for organs 3D visualization and operation simulation. The precise of the segmented object is critical for doctor´s diagnosis and disease treatment. The SeeThrough application is medical image process that includes image segmentation, 3D object visualization, furthermore aiding diagnosis.
    • Algorithm : Segmentation methods are usually divided into two region-based and edge-based methods. In region-based methods an algorithm usually searches for connected regions of pixels with some similar feature such as brightness, texture pattern, etc. These algorithms work in the following way: The first image is in some way divided into regions. Then for each region similarity among pixels is checked. If similarity is below some threshold, region is divided into smaller regions. In the next step neighbouring regions with similar features are merged into a new bigger region.These two steps are repeated until there is no more splitting or merging. The problem in this approach is to determine exact borders of objects because regions are not necessary split on natural borders of the object.

      Alternative approach is edge-based.In this approach an algorithm searches for pixels with high gradient value which are usually edge pixels and then tries to connect them to form a curve which represents a boundary of the object. A difficult problem here is how to connect high gradient pixels because in real images they are usually not neighbours. Another problem is noise. Since a gradient operator is of a high pass nature, and the noise is usually also in high frequencies it can sometimes create false edge pixels.

      In this application,the region-based method and edge-based method will be combined to realize the general medical image segmentation, thereafter perform 3d object visualization and volume measurement, and so on.

    • Screen Shot :

      Figure 1 : SeeThrough

  • 3D-Generator
    • Introduction : The application is used for the dentist diagnosis. During the check up procedure, the dentist gets a serial of cross-section image of patient. In every image, the teeth or gum area is arched shape. However, the dentist wants to warp the arc shape to line shape, then construct 3D image, so that he can observe the surrounding area in one plane. Our application can fulfill this requirement that perform deformation the series of images, and also visualize the result as 3D
    • Algorithm : The deformation is performed on 2D image. Then using these deformed images we construct 3D image. Our method is based on the work of Thaddeus Thaddeus Beier and Shawn Neely. Feature-based image metamorphosis.Computer Graphics (SIGGRAPH'92 Proceedings). Vol. 26, July, 1992. pp. 35-42. It is local morphing for that the morphing is based upon fields of influence surrounding the control elements. Using this method,it is easy to define the feature primitives. And the algorithm is intuitive and easy to realize by programming.
    • Screen Shot :

      Figure 2: 3D-Generator


Other Links

... to our Research
              Security Engineering - Learning & Knowledge Tech - Design Thinking - former
... to our Teaching
              Tele-Lectures - MOOCs - Labs - Systems 
... to our Publications
              Books - Journals - Conference-Papers - Patents
... and to our Annual Reports.