36 AAPM Grand Challenge for people to study a challenging problem with a common dataset,” he points out. “It’s an opportunity for people to establish the performance of their own algorithms. That’s very important if they want to develop novel algorithms.” The challenge seeks to explore improvements in parameter reconstruction beyond the current clinical standard by leveraging advancements in deep learning and classical reconstruction algorithms. Participants will be provided with a pre-prepared sample code to understand the baseline level upon which to improve. For those participants who are not MRI researchers or familiar with MRI principles, Xun says understanding the problem at hand is vital for optimal performance. “In reconstruction, we talk about the forward problem going from the solution to the measurement,” he points out. “The actual challenge is the inverse problem, going from the measurement to the solution. Participants have to understand the forward problem carefully.” Karen agrees: “For people not from the medical physics field, it’s good to know the actual clinical question and its importance. We get a lot of participants for challenges that are in no way, shape, or form from medical physics or even medical imaging. They come from computer vision or completely unrelated fields. It’s a way to bring more expertise into the field.” “It’s an opportunity for people to establish the performance of their own algorithms. That’s very important if they want to develop novel algorithms!”
RkJQdWJsaXNoZXIy NTc3NzU=