Mr Abbas Khan Rayabat KhanComputer Science PHD StudentEmail: a.rayabatkhan@qmul.ac.ukProfilePerformanceProfileProject: AI-based Cardiac Image Computing Supervision: Maritin Benning (UCL), Dr Caroline Roney and Prof Greg Slabaugh Why do a PhD?After my Master's degree, I could work in industry or pursue a Ph.D., and I chose the latter. I was enthusiastic about delving even deeper into the AI field and spending more time exploring different directions in AI. Moreover, following the recent trend in the market, I found that having a Ph.D. can open up additional career opportunities, particularly in academia and research. What are you researching and what led you work in this field of research?I am working on AI-based Cardiac Image Computing, more specifically, segmentation. This aims to help cardiologists enhance diagnostic accuracy, monitor disease progression, and enable personalized medicine. Following my father's passing due to a heart stroke, I resolved to dedicate myself to combating cardiac disease. Equipped with an engineering degree and a master's in precision agriculture, I initially questioned whether I could contribute meaningfully to this cause. However, my excitement soared upon discovering the CTD program at Queen Mary University of London's Faculty of Science and Engineering, which focused on AI-based cardiac image computing. Despite having numerous alternatives, I pursued this opportunity vigorously, as it aligned perfectly with my passion and purpose. Why did you choose DERI?DERI is a prestigious institution renowned for its position at the forefront of cutting-edge multidisciplinary research. I am immensely fortunate to have the opportunity to contribute to DERI's rich academic environment. The institute boasts state-of-the-art technologies, ensuring all users can access the tools necessary to advance innovative research endeavours. Beyond its educational resources, DERI fosters a vibrant social atmosphere, offering occasions to celebrate various events and connect with colleagues outside the realm of research. One of the standout benefits of being affiliated with DERI is its access to the Andrena cluster, a pinnacle of advanced computing infrastructure. Also situated centrally amidst numerous tube and bus stations, DERI's location in East London facilitates convenient commuting for researchers, making accessibility an essential aspect of its appeal. Being part of DERI entails many advantages that enrich academic pursuits and personal experiences, embodying an environment where excellence thrives. Have you any key tips on keeping motivated throughout a PhD?Meeting your supervisory team regularly is crucial for maintaining motivation during your PhD journey. In twenty minutes, your supervisors can distil their twenty years of experience into valuable insights. Keep informing them about your research goals and seek their support when facing challenges. Additionally, being kind to yourself is essential, as staying focused on your goals and remembering the initial passion that drove you to embark on this journey. What online resources do you recommend?To stay abreast of current research trends, I actively follow professionals and thought leaders in my field on LinkedIn. Additionally, I utilize various platforms like ChatGPT, Stack Overflow, Medium, GitHub, and ChatPDF to seek guidance, solve coding issues, and deepen my understanding of concepts related to my work. These platforms offer a wealth of resources and insights that help me stay informed and continuously improve my skills in my area of expertise. What is the most intriguing puzzle that your research has revealed so far?The core focus of my Ph.D. research revolves around employing compositional or break-down approaches to address intricate problems. The underlying principle involves adopting a hierarchical perspective, where subsequent steps capitalize on the insights from preceding tasks. Our investigations have discovered that such sequential methodologies consistently yield superior results compared to direct methods. Moreover, this approach enables us to effectively eliminate outlier predictions and develop algorithms explicitly tailored for cardiac image segmentation, enhancing our solutions' precision and focus. Connect with Abbas :GithubGoogle Scholar Highlights:Two papers led by Abbas have been accepted for publication at WACV2025: CAMS: Convolution and Attention-Free Mamba-based Cardiac Image Segmentation, introduces a novel architecture based on Mamba for cardiac image segmentation. Compositional Segmentation of Cardiac Images Leveraging Metadata, provides global-to-local anatomical segmentation that leverages metadata such as field strength and machine parameters. Paper led by Abbas was also presented at the International Symposium on Biomedical Imaging, held in Athens in May 2024: [2402.09156] Crop and Couple: cardiac image segmentation using interlinked specialist networks (arxiv.org) Education Background information: Master’s Degree in Electronics and Information Engineering from Jeonbuk National University (JBNU), Jeonju, Republic of Korea, (2019-2021) with research focused on fundamental Deep Learning and applications of Computer Vision to Healthcare and Precision Agriculture. I was supervised by Professor Hyongsuk Kim and Professor Yongchae Jeong , as well as working as a Research Assistant at Core Research Institute of Intelligent Robots. Bachelor’s Degree in Electrical and Electronics Engineering with Magna Cum Laude from (2014-2018). I worked on Retinal layers segmentation in my final year and was supervised by Dr. Taimur Hassan . ResearchPerformanceHonors and Awards, Scholarships and Internships. MICCAI 2022 LAScarQS Challenge Best Paper Award. Brain Korea (BK21) Scholarship Bahria University Advanced Merit Scholarship CERN, Switzerland “Conseil Européen pour la Recherche Nucléaire” Summer Program (2018 Türkiye Belediyeler Birliği – Çankaya/Ankara, Turkey, LocalInternational Internship Program Research Intern 01/06/2022 – Present. Research Intern from June 2022 to date: Keen AI, Software Company in Birmingham, United KingdomDeveloping AI-based approaches to segment steelwork and rust on images of transmission towers. The project aims to precisely predict the steelwork on various backgrounds and then estimate the proportion of the rust. The images are collected through drones, helicopters, and simulations, the project won an AIM Asset Management Excellence Award in 2023. Research Assistant from Dec 2023 to date: NIHR Barts Biomedical Research Centre, London, United KingdomI am working on innovating in drug and device development for the cardiovascular digital twin concept to digitally model acute coronary syndromes, valvular heart disease, and hypertension. For each workstream, I am working on the ML algorithm development side. For valvular heart disease, I developed the precise segmentation of the Aorta; for the other two studies, we are developing time-series solutions for continuous monitoring of the patients based on sensor data and demographics.