CT Visualization is a set of techniques used to display a 2D projection of 3D discretely sampled data set from Computed Tomography (CT). Typically, a CT scanner produces a group of 2D sliced images well ordered in space volume, representing density of the scanned object. With the rapid development of CT, MRI and other biomedical imaging techniques, many diseases are able to be detected early with non-invasive examination. However, such diagnosis, more or less, depends on doctor’s judgment. Hence, faster rendering, higher resolution, better arranged display and other improvements to represent CT Data, will be very beneficial in disease diagnosis. By computer graphics and scientific visualization, the investigator is now able to quickly obtain images with explicit geometric structure as well as tissue difference of certain part of the body. In pass few decades, as incredible booming growth in Graphics Processing Unit (GPU) industry, a trend that to applied advanced GPU techniques in visualization has been developed. Computing and rendering in high-performance GPU instead of CPU results in better quality of the output images, which is significant to improve the accuracy of doctor’s diagnosis.
In my final project, I am going to propose a method using OpenGL and CUDA to transform density CT data from 2D gray-level sliced image sequence into 3D volume. The first thing I am going to do is to render the volume space in openGL, adding exploration controls, e.g., rotation, zooming, shifting etc. And then add different rendering methods, for example different to compare their results and performance. At last, I will try to colorize the image to distinguish different tissues and organs as the image showed below. In addition, I will try to read DICOM data, the general data format for medical images, and implement my program on android device with openGL ES.
Here is the pdf link:
Final Project Brief Overview