Andrew Top wins innovation award for paper on 3D biomedical images
Congratulations to Andrew Top for receiving the 2013 WAGS Innovation in Technology Award for his master's thesis: Automated confidence-based user guidance for increasing efficiency in interactive 3D image segmentation. He'll be going to the annual WAGS conference in Tucson, Arizona, this March to accept the award.
It wasn't his first award. Not only did he get A+ marks in all of his courses; he also received a Pacific Century Graduate Scholarship, a Helmut Eppich Graduate Scholarship and an SFU Graduate Fellowship in the course of his graduate studies.
His graduate research addressed a critical need for doctors, radiologists and researchers to separate regions of interest in 3D images produced by Computed Tomography (CT) scanning, Magnetic Resonance Imaging (MRI) and other modalities. For example, in order to aim the radiation in radiation therapy, an oncologist must first identify the exact location of the tumor in a CT scan of the patient. The process of identifying or categorizing regions of interest in an image is called image segmentation.
Andrew's thesis focused on improving the quality and speed of interactive 3D image segmentation methods by introducing the idea of having the computer intelligently guide the user throughout the process. The technique was first published at the Medical Computer Vision workshop held in conjunction with the international Medical Image Computing and Computer Assisted Intervention (MICCAI) conference in 2010, where the research team named it Spotlight.
His results showed that the use of Spotlight reduced the amount of user time required to segment a human liver (among other examples) by 35% compared to segmenting without the use of Spotlight. Subsequent improvements after the paper was submitted have seen that speed-up rise to 60%.
A generalization of his initial research was then published at MICCAI 2011, and received the additional honour of an invitation to deliver a presentation to a crowd of over five hundred medical image analysis professionals from around the world — it's a privilege given to only 4% of the papers submitted.
He found his way to academia after working in the video games industry and discovering that he wanted to apply more math to his work. When he was contacted by Dr. Ghassan Hamarneh about doing research at the Medical Image Analysis Lab (MIAL) he thought that the position sounded like the perfect mix of computer science and math, and happily joined the team.
He says, "witnessing the public release of TurtleSeg, a software application for interactive segmentation that incorporates a number of medial image analysis methods — including that described in my thesis — was incredibly rewarding for me. The software has been downloaded thousands of times all over the world (see map of downloads) and it has been a great pleasure to watch the software grow in popularity. It is clear also that the field of automated guidance in 3D image segmentation is growing as there have been many papers published since our first publication on the subject in 2009."