2025, Jul Our work on recurrent medical image registration is published on Medical Image Analysis. Formulating registration as a meta-learning problem boosts training data efficiency by up to 20x. [paper] [code]
2025, Jun Our study on 'slim' nnU-Net is accepted by Medical Imaging with Deep Learning (MIDL) 2025, Salt Lake City. [paper] [code]
2024, Nov Qian gave an invited talk at the 4th Annual Meeting of the Dutch Inverse Problems Community, Delft.
2024, Oct Qian co-organized the 15th STACOM Workshop at MICCAI 2024, Marrakesh, Morocco.
2024, Oct Our recent work on deep-learning-based groupwise registration for cardiac T1 mapping is accepted by MICCAI 2024. Check out how we blend PCA and CNN, the old and new in machine learning, to tackle the challenging task of heart motion correction under varying contrast. [paper]
2024, Oct Is high uncertainty (low confidence) always a bad thing? In our MICCAI paper, we show how we can leverage uncertainty to our advantage in solving difficult problems such as right ventricle segmentation from 2D+t cine MRI, a persisting challenge especially at the right ventricular outflow tract (RVOT). In addition, we share the updated RVOT annotations of the ACDC dataset for future studies of RV segmentation. [paper] [data]
2024, July Three short papers were accepted by Medical Imaging with Deep Learning (MIDL) 2024, Paris. Congratulations to our BEP and MEP students and thanks for your excellent contributions!
2024, May We launched the first AI course in the Applied Physics master program, together with the team of Dr. Eliška Greplová, Department of Quantum Nanoscience.
2024, Apr Our project on fair data for medical imaging is funded by the National Growth Fund (NGF) - AiNed XS Europa.
2024, Apr We received an Amazon Research Award on generative AI.
2024, Feb Yunjie's paper on MRI translation is accepted by the Journal of Machine Learning for Biomedical Imaging. Check out how neural fields overcome the spectral bias of CNNs in image generation. [paper] [code]
2023, Dec Qian gave an invited talk at the Special Seminar "ML and AI for Medical Applications", Johann Radon Institute for Computational and Applied Mathematics (RICAM), Austrian Academy of Sciences, Linz, Austria.
2023, Oct Another interview by Delta, the independent journalist platform of TU Delft, on our Radiology AI research. Qian discussed with Kim our approach to AI trustworthiness (yes, by the seemingly contradictory uncertainty).
2023, Oct Qian served on the MICCAI 2023 Best Paper & Young Scientist Award Committee. This is the first time that MICCAI gave the Best Paper Award besides the Young Scientist Award. The first MICCAI Best Paper Award goes to Alvaro Gonzalez Jimenez for his work on Robust T-Loss. Long live the Student-t distribution!
2023, Oct Yidong won a 3rd prize in the highly competitive Cardiac MRI Reconstruction Challenge (CMRxRecon). Congratulations! [paper] [code]
2023, Oct Qian delivered the Keynote speech at STACOM, MICCAI 2023, Vancouver.
2023, Sep Our project on uncertainty of deep learning was supported by the NVIDIA Applied Research Accelerator Program. Our computation was substantially fueled by the high-performance computing of NVIDIA LaunchPad!
2023, Sep Qian co-organizes the 2023 Netherlands Conference on Computer Vision (NCCV). Thanks Yi for being our website chair!
2023, Jun Qian gave an invited talk at the Women in Data Science Event, University of Twente.
2022, Nov Qian is featured as a Delft Health Initiative (DHI) Rising Star. Thanks Merel for the engaging interview!
2022, Oct Qian served as a jury member for the Young Medical Delta (Leiden-Delft-Rotterdam) Thesis Awards, representing TU Delft. It was great fun to hear about all the out-of-the-box projects!
2022, May Workshops are traditions and treasures of MICCAI. Qian serves as the Workshop Chair for MICCAI 2022, Singapore.
2022, May Ever wonder when the amazing, prize-sweeping nnU-Net could still fail? Check out Yidong's MICCAI paper on nnU-Net's uncertainty. [paper] [code] [blog]
2022, May Qian gave an invited talk at the EACVI/SCMR Joint Summit on Artificial Intelligence in CMR, London: "AI for CMR postprocessing - Is the problem solved?" In our panel discussion, we tried hard to identify unsolved problems, with cardiologists, radiologists, and MR physicists. (Perhaps we should blame the U-Net for taking our problems away?)
2022, May Xinrui's visualization work SpaceMAP was accepted as Spotlight by the International Conference on Machine Learning (ICML). Check out the beautiful Word Map of World Map (Fig. 10) and Prime Number Divisibility (Fig. 11). Congratulations Xinrui, also on the ICML Student Participation Grant! [paper] [code]
2021, Nov Our curiosity project on visualizing gender differences received seed funding from the Delft Health Initiative.
2021, Oct Our visualization work, Deep Recursive Embedding, is accepted by IEEE Transactions on Visualization and Computer Graphics. This joint work with Prof. Boudewijn Lelieveldt (LUMC) shows how neural networks can learn to visualize high-dimensional data in the same (or better) ways as t-SNE and UMAP, while being more scalable. Check out the fun movie of MNIST visualization - without any supervision! [paper] [code]
2021, Jul Qian served as Program Committee (PC) member for the Association for the Advancement of Artificial Intelligence (AAAI).
2021, Mar Qian received the Editor’s Recognition Award, with Special Distinction, from Prof. Charles E. Kahn, Jr., Editor-in-Chief of Radiology: Artificial Intelligence.
2021, Mar Our new lab is funded by the TU Delft AI Initiative and the Department of ImPhys.