Computer Vision News - April 2023

51 Edited-MRS Reconstruction still stuck in the research setting . Rodrigo is exploring this problem as a part of his thesis, aiming to break down that clinical barrier. “ I hope this challenge attracts more people from the medical imaging community and brings greater awareness to edited MRS and how it can be improved with deep learning, ” he tells us. “ I’d like to establish an initial benchmark on how that improvement could be performed specifically for accelerating MRS. Hopefully, we will find it reaching that clinical and research barrier in the next few years. ” Hanna concludes: “ I hope the challenge increases the confidence between science and machine learning . There’s still a question: Can machine learning improve, and can we trust it? Doing more of these challenges and seeing the outcomes will help improve that confidence that machine learning is a way to go for certain things! ” public MRI dataset – on hand to advise and guide them through it. “ The University of Calgary is pretty good at MRI and MRS – specifically, their combination, ” Rodrigo reveals. “ We have many colleagues who work on projects related to MRI, MRS, and machine learning. It’s a big field here. ” Hanna, currently transferring to a PhD at the university, agrees: “ It’s a pretty big imaging hub here, and our supervisor came to us with this great opportunity to propose a challenge at ISBI . Many articles on MRS and deep learning have emerged recently, but I was amazed that I could only find one MRS challenge in the last 10 years . If you can’t find a challenge, why not start one and drive that innovation forward? I think now is the right time to push that boundary. ” So far, there has been a greater focus on MRI because it has a more clinical application than edited MRS, which is