Jingxian Sun BEng, MSc, Presenter: Nothing to Disclose

Qiang Meng, Abstract Co-Author: Nothing to Disclose

Jing Qin, Abstract Co-Author: Nothing to Disclose

Pheng Ann Heng PhD, Abstract Co-Author: Nothing to Disclose

Visible human images have unique value for medical education and research because they has higher resolution than CT, MR or Ultrasound images, and can exhibit richer anatomical details. However, these images are usually obtained along a constant direction and hence users cannot explore anatomical structures at arbitrary positions and angles, which is quite important for surgical path planning. Moreover, the original images lack of labels. To solve these two problems, we build an interactive learning system to display and label an anatomical image obtained from arbitrary positions and angles in real time. In the system, users use a tracker to locate a position on a 3D printed plastic human model, and the corresponding sectional image as well as its label information are calculated and displayed. In order to achieve real-time performance, we use GPU to accelerate the visualization and labeling processes.

We invited two groups of volunteers from medical school to evaluate the interactivity and usability of our system. One group uses a system with keyboard and mouse as input devices while the other group uses our system. Both groups are asked to find the locations of specified tissues. The results showed the second group completed the task averagely three times faster than the first group. Moreover, we found our system is easy to use that users can skillfully operate it with a few instructions.

Displaying labels in real time is an important feature of our system. However, because the anatomical structures in the images are complicated, it is extremely difficult to identify and segment all tissues in the images. As a pre-processing step, we manually segmented and labeled about 1000 tissues in the original images. We will segment more organs and tissues to provide users a more complete learning environment.

We developed an easy-to-use anatomy learning system with a 3D printed human body and a 6DoF tracker. Slice images and labels are real-time calculated and displayed as users move the tracker on the plastic model at arbitrary positions and angles. This system can be used in the human anatomy education and research as well as other applications such as surgical planning. 

http://abstract.rsna.org/uploads/2014/14005912/14005912_fbhi.jpg

Sun, J, Meng, Q, Qin, J, Heng, P, An Interactive Sectional Anatomy Learning System Based on Chinese Visible Human Dataset.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14005912.html