Surgical Simulations, Asian Fonts, and Special Effects for Movies
Sarah Frisken, PhD, joined the Department of Computer Science in 2005. She works in computer graphics and modeling and is currently involved in three main research areas: medical imaging for applications such as anatomy education, surgical simulation, and computer-assisted surgery; digital font representation; and 3D computer-based design, sculpting and modeling. Many of Frisken’s projects are tied together by her work with Adaptively Sampled Distance Fields, a new way of representing complex shapes on the computer that provides fine graphic detail with more efficient use of computer resources.
After earning her PhD in electrical and computer engineering from Carnegie Mellon University in Pittsburgh, Frisken did a postdoctoral fellowship in the Neuman Biomechanics Laboratory at Massachusetts Institute of Technology, where she worked on automatic segmentation of magnetic resonance images. She returned to Carnegie Mellon, where she co-founded the Center for Medical Robotics and Computer-Assisted Surgery, and conducted research in medical image processing, surgical simulation, and computer-assisted surgery. She also worked at Mitsubishi Electric Research Laboratories for 11 years, where she co-developed Adaptively Sampled Distance Fields and applied them to research areas as varied as digital font representation and simulation of arthroscopic knee surgery.
The Tufts Digital Anatomy Project is currently Frisken’s broadest endeavor, involving the departments of Anatomy and Cellular Biology, Biomedical Engineering, Computer Science, Mathematics, Pathology, and Radiology. The main goal of this project is to extend the value of the Anatomical Gift Program beyond a one-time dissection by using full-body computed tomography (CT) to scan cadavers that have been donated to the program. “Students will be able to preview the anatomy of a specific cadaver prior to dissection, compare their lab findings to the digital images, and review anatomy long after dissections have been completed,” says Frisken. The Digital Anatomy Project may also afford anatomy training to non-medical students such as biomedical engineers for whom medical gross anatomy courses are not available.
“We also anticipate that this data will become a valuable national or even international resource,” says Frisken. CT image data is currently hand processed to create an anatomical computer model. Frisken’s group is designing software algorithms for processing the data that will allow researchers to build 3D models from the CT scans more automatically. These models could then be used not only for visualizing 3D anatomy but also for performing quantitative studies (such as measuring bone density or fat distribution) and for building functional models (of a beating heart, for example, or of joint dynamics) that would be useful for simulating and planning surgeries.
Frisken’s second major interest is in digital font representation. The system she worked on while at Mitsubishi Electric Research Laboratories is now the main font rendering engine of Adobe’s Flash technology (formerly known as Macromedia Flash), which is estimated to have over 500 million users. She is currently working with a graduate student on a new way of representing and rendering Asian fonts that is compact, expressive, and provides scalable type of high quality even at small point sizes. These features would make the new representation particularly attractive for use on today’s smaller electronic devices.
Computer-based free-form design and modeling, which includes 3D digital sculpting and concept design, is Frisken’s third major interest. Her work in digital sculpting, which is also based on Adaptively Sampled Distance Fields, has attracted the interest of a movie industry technical director who uses digital effects. “He’s looking for a better way to create realistic computer models of terrain for special effects because current systems have a hard time modeling large terrains with enough detail for close-up shots as well as terrains with caves and steep cliffs,” says Frisken. “He’s really excited about our digital sculpting system; we’ve begun working with their team to explore building a terrain modeler based on our technology.”
Tufts’ atmosphere of easy collaboration was a big draw for Frisken. She has already made strong links throughout the university and has several active collaborations. She would welcome a collaborator interested in working on an anatomy curriculum that incorporates imaging technology and 3D computer modeling. She would also welcome help in evaluating how the new curriculum affected the students’ understanding of anatomy.
For more information, please go to http://www.cs.tufts.edu/~frisken/frisken.html.