{"id":"arxiv/2610.27869","url":"https://arxiv.org/abs/2610.27869","tags":["math.NA"],"score":0,"title":"Inverse domain uncertainty quantification in hemodynamics: sampling virtual patient cohorts under constraints","domain":"math.NA","source":"arxiv","authors":["Francesco Romor","Jose Rodrigo Rojo-Garcia","André-Alexander Zepernick","Ana Djurdjevac","Alfonso Caiazzo"],"reviews":0,"subject":"math.NA","summary":"We consider the following inverse domain uncertainty quantification problem: given a prescribed value or measurement of a hemodynamic biomarker, characterize the distribution of vascular geometries compatible with it. The problem is formulated as a Bayesian inverse problem in which a high-dimensional latent parameter encodes the cardiovascular shape and accounts for inter-patient variability. The parameter-to-observable map composes a conditional LDDMM stochastic interpolant generating the aortic surface with a high-performance numerical finite element solver for the unsteady incompressible…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-10-20","reactions":{}}