Using two-dimensional edge detection to produce three-dimensional medical prototypes from MRI data

Currently, medical practitioners can export medical resonance imaging (MRI) files to specialized software technologies to create a rapid prototype of bone or soft tissue. This specialized software may not be readily available to smaller medical practices; the purpose of this undergraduate research project (sponsored in part by the NASA Space Grant Internship program) was to develop a process more suitable for use with conventional computation engineering tools. Initially, we developed a process which examined two-dimensional slice images from the MRI data (provided in the form of an audio video interleave (AVI) file). A simple edge threshold detection method identified bone boundaries. However, this method proved troublesome for slice data associated with interior surfaces of cavity features, such as that represented by the human skull. This experience led to the development of a three-step process. The Canny edge-detection method was used as a preprocessor on the slice-images, with the purpose of enhancing the detection of bone boundaries. The second step involved refining the bone boundary locations into specific X-Y coordinate positions through the use of a Matlab computer program. These sets of slice data points were then assembled as a three-dimensional point cloud and imported into a computer-aided design package (SolidWorks). This CAD software enables the user to further manipulate the model into an STL file format, suitable for reproduction on a rapid prototyping machine.