Jaw Viewer: Displaying jaw movement for medical diagnosis

Introduction
As its name indicates, the Clinic of Masticatory Disorders of the Zurich University offers treatment for facial pain and mandibular joint problems. For better diagnosis, the Clinic has developed a proprietary 3D camera system (Optis) to record the patient's mastication movement. They also have the tools to extract the individual bones from the MRI images of a patient's jaw and store them as Stereo Lithography Files. In another proprietary software called TMJViewer, they can combine both of these resources to display the movement in a 3D animation on a computer screen and use that to give an elaborate analysis as well as feedback. But the current process required to achieve this goal does not only have lots of stages and different systems involved, but is also tedious and above all, error-prone. All required files come from different sources and must be manually processed with help of different applications until they can be used properly. Besides, this process does not allow a display of the movement in real time, which can lead to frustration on both ends since direct communication and feedback is not possible. A wrong recording or unnatural movement of the patient can lead to unusable data and require multiple visits in an attempt to be corrected. Thus, the idea to develop a new application with the goal of unifying all these processes and supporting real time movement display, using the graphics API OpenGL, was created.

Approach / Technologies
As a first step, we were required to get familiar with OpenGL as well as various aspects of graphics rendering. To achieve this, we consumed some literature on the topic and developed a C++ prototype based on a given tutorial. We then used that prototype as a base for our project and extended its functionality based on the requirements. Our software needed to be able to import both the anatomical objects (as STL files) and the recorded movement (as MVM files, a proprietary format). The objects then had to be displayed and the movement extracted into transformation matrices, that can be interpreted by the graphical components to set selected objects in motion. We then refactored the application before integrating it into a new WindowsForms application to provide an appropriate graphical user interface for its configuration. Due to a lack of information and unclear directions, the displayed movement ended up not being completely accurate and a lot of reverse engineering was necessary to reach the resulting version of our software solution.

Result

This project resulted in an application which can unfortunately not be used by the Clinic because of the explained inaccuracy, but which serves as a good basis for future projects. The documentation serves as introduction to OpenGL and explains among other topics, the mathematical foundations needed to understand the complex domain and implement the calculations necessary for displaying motions on screen.