[Defense] Novel Deep Learning Approaches for LV MRI Analysis and Simulations of Ferric Applicators for Therapeutic Interventions
Wednesday, November 3, 2021
12:30 pm - 2:00 pm
will defend his dissertation
Novel Deep Learning Approaches for LV MRI Analysis and Simulations of Ferric Applicators for Therapeutic Interventions
Left ventricle (LV) segmentation is critical for clinical quantification and diagnosis of cardiac images. In this work, we propose three novel deep learning architectures called BNU-Net, LNU-Net and IBU-Net for left ventricle segmentation from short-axis cine MRI images. BNU-Net is the batch normalized (BN) U-Net, LNU-Net is the layer normalized (LN) U-Net, and IBU-Net is the instance-batch normalized (IB) U-Net. The architectures of BNU-Net, LNU-Net and IBU-Net have an encoding path for feature extraction and a decoding path that enables precise localization. BNU-Net, LNU-Net and IBU-Net have left ventricle segmentation methods: BNU-Net employs batch normalization to the results of each convolutional layer, LNU-Net applies layer normalization in each convolutional block and is based on an exponential linear unit (ELU), while IBU-Net incorporates instance and batch normalization together in the first convolutional block and applies an ELU and passes its result to the next layer.
Our method incorporates affine transformations and elastic deformations for image data processing. Our dataset that contains 805 MRI images regarding the left ventricle from 45 patients is used for evaluation. The experimental results reveal that our approach accomplishes comparable or better performance than other state-of-the-art approaches in terms of the dice coefficient and the average perpendicular distance. Magnetic Resonance Imaging (MRI) is a well-established modality for pre-operative planning and is explored for intra-operative guidance of procedures such as intravascular interventions. We simulated a computational platform for preoperative planning and modeling of MRI-powered applicators inside blood vessels. This platform was implemented as a two-way data and command pipeline that links the MRI scanner, the computational core and the operator. The platform first processes multi-slice MR data to extract the vascular bed and then fits a virtual corridor inside the vessel. This corridor serves as a virtual fixture (VF), a forbidden region for the applicators to avoid vessel perforation or collision. The geometric features of the vessel centerline, the VF and MRI safety compliance (dB/dt, max available gradient) are then used to generate magnetic field gradient wave forms. Different blood flow profiles can be user-selected, and those parameters used for modeling the applicator’s maneuvering.
12:30PM - 2:00PM CT
Online via MSFT Teams (link TBA)
Dr. Nikolaos Tsekos, dissertation advisor
Faculty, students and the general public are invited.