Title: Deep Learning in Biomedical Imaging
Presenter(s): Ngan Le
Date: December 1, 2021
Medical image analysis using deep learning has recently been prevalent, showing great performance for various downstream tasks including medical image segmentation and its sibling, volumetric image segmentation. Particularly, a typical volumetric segmentation network strongly relies on a voxel grid representation which treats volumetric data as a stack of individual voxel ‘slices’, which allows learning to segment a voxel grid to be as straightforward as extending existing image-based segmentation networks to the 3D domain. However, using a voxel grid representation requires a large memory footprint, expensive test-time and limiting the scalability of the solutions. Furthermore, it is arguable that in traditional CNNs, its pooling layer tends to discard important information such as positions and CNNs are sensitive to rotation and affine transformation.
In this presentation, we present Point-Unet (accepted by MICCAI 2021), a novel method that incorporates the efficiency of deep learning with 3D point clouds into volumetric segmentation. We also introduce 3D-UCaps, a 3D voxel-based Capsule network for medical volumetric image segmentation (MICCAI 2021).
Dr. Ngan Le is the director of Artificial Intelligence and Computer Vision lab and an Assistant Professor in the Department of Computer Science & Computer Engineering at University of Arkansas. She was a research associate in the Department of Electrical and Computer Engineering (ECE) at Carnegie Mellon University (CMU) in 2018-2019. She received the Ph.D degree in ECE at CMU in 2018, ECE Master degree at CMU in 2015, CS Master Degree at University of Science, Vietnam in 2009 and CS Bachelor degree at University of Science, Vietnam in 2005.
Her current research interests focus on Image Understanding, Video Understanding, Computer Vision, Robotics, Artificial Intelligence (Machine Learning, Deep Learning, Reinforcement Learning), Biomedical Imaging, SingleCell-RNA.
Dr. Le is currently a guest editor of Scene Understanding in Autonomous (Frontier) and Artificial intelligence in Biomedicine and Healthcare (MDPI). She co-organized the Deep Reinforcement Learning Tutorial for Medical Imaging at MICCAI 2018, Medical Image Learning with Less Labels and Imperfect Data workshop at MICCAI 2019, 2020, 2021, Visual Detection, Recognition and Prediction at Altitude and Range at ICCV 2021. Dr. Le is instructor lead of the Google Machine Learning Bootcamp 2021. Her publications appear in the top-tier conferences including CVPR, MICCAI, ICCV, SPIE, IJCV, ICIP etc, and premier journals including IJCV, JESA, TIP, PR, JDSP, TIFS, etc. She has co-authored 72+ journals, conference papers, and book chapters, 9+ patents and inventions. She has served as a reviewer for 10+ top-tier conferences and journals, including TPAMI, AAAI, CVPR, NIPS, ICCV, ECCV, MICCAI, TIP, PR, TAI, IVC, etc
Learn more about Dr. Le’s research by visiting https://www.nganle.net.