Student Chapter
Student Chapter
This chapter is a nationwide chapter to help promote computational mechanics to young researchers and improve student engagement and professional development within our community.
USACM Student Chapter Seminar Series
October 22, 2025; 3PM EDT
Join via Zoom: https://us06web.zoom.us/j/82464478256?pwd=ZMkJVFdjMJzadgnVWFPqsdUSs4qTaY.1
Speaker: Pan Du, University of Notre Dame
Title: A hierarchical NURBS-based generative model for aortic geometry synthesis and controllable editing
Abstract: Cardiovascular diseases (CVDs) remain the leading cause of mortality worldwide, encompassing a wide spectrum of pathologies such as aortic aneurysms and coarctation. Characterizing vascular anatomy is therefore essential for diagnosis, prognosis, and treatment planning. Advances in medical imaging, including MRI and CT, now enable high-resolution reconstructions that underpin quantitative analysis and computational fluid dynamics (CFD) simulations of patient-specific hemodynamics. These digital models have facilitated in silico experimentation, risk assessment, and even 3D-printed phantoms for surgical planning and device testing. Beyond individual cases, large cohorts of vascular geometries are crucial for population-level statistics and pathology analysis, as well as for training deep-learning-based CFD surrogate models. This requires generative tools capable of producing large numbers of realistic, simulation-ready anatomies.
To address these challenges, we present HUG-VAS, a Hierarchical NURBS Generative framework for Vascular models that unifies NURBS-based 3D encoding with diffusion-based generative modeling to synthesize fine-grained, CFD-ready aortic anatomies. HUG-VAS also enables training-free, zero-shot conditional synthesis through diffusion posterior sampling (DPS) from image-derived prompts such as sparse 3D points or partial contours, supporting robust reconstruction and interactive editing even under degraded imaging. With sufficient image guidance, the uncertainty in the generated ensemble collapses toward the ground truth, introducing a new paradigm for vascular segmentation: semi-automatic segmentation. Unlike existing workflows that are either manual (accurate but slow) or fully automatic (fast but prone to artifacts), this hybrid diffusion-guided approach ensures efficient and anatomically consistent geometry extraction with flexible post-editing capability. Altogether, HUG-VAS establishes a unified path from limited imaging to large-scale, simulation-ready vascular digital twins—bridging data generation and segmentation within a single generative framework.
Bio: Pan Du is a Ph.D. candidate in Aerospace and Mechanical Engineering at the University of Notre Dame, advised by Prof. Jian-Xun Wang. His research lies at the intersection of computational physics and artificial intelligence, with a focus on advancing biomedical and health sciences. He develops deep learning, generative AI, probabilistic inference, and differentiable solver frameworks for tasks such as medical image segmentation, patient-specific vascular geometry generation, and hemodynamic flow prediction. These efforts aim to create digital twins of the cardiovascular system, enabling improved diagnosis, risk assessment, and treatment planning for patients with complex vascular diseases.
Past Seminars
Past Recordings of the Student Chapter Seminar Series can be found here.
Past Student Chapter Seminar Information can be found here.
Student Chapter News
There is currently no student chapter news.
FAQ
Q: Are there additional dues required to join the USACM student chapter?
A: No additional dues are required.
Q: Who is eligible to join the USACM student chapter?
A: All people interested in computational mechanics are welcome to join. However, only USACM student members are eligible for student events.
Additionally, USACM student members who have graduated within 2 years are also eligible to participate in student events.
Q: How do I join the USACM student chapter?
A: Join us on our official discord server to keep up to date with our activities.
Q: How do I become a USACM student member?
A: Fill out the membership form.
Organization
January 2025 - December 2026
Mentors
Jessica Zhang, Carnegie Mellon University
John Evans, University of Colorado Boulder
David Littlewood, Sandia National Laboratories
Patrick Diehl, Los Alamos National Laboratory
Osama Raisuddin, Rensselaer Polytechnic Institute
Student Leadership
Chair: Yanrong Xiao, Yale University
Vice-Chair: Joan Ejeta, North Carolina A&T State University
Member-at-Large: Joseph Amponsah, Iowa State University
Elaheh Mehdizadeh, University of Pittsburgh
Dila Kandel, University of Utah
Executive Members
Kishore Appunhi Nair, Johns Hopkins University
Gargeya Bhamidipati, Johns Hopkins University
Xuan Hu, University of California, Berkeley
Ozge Ozbayram, Johns Hopkins University