Abstract
Reconstructing geometrically consistent 3D cardiac surfaces from sparse short-axis (SAX) cine MRI is challenging because of discontinuous long-axis sampling and ambiguous inter-slice geometry. Existing methods typically learn in mesh vertex spaces or generic 3D coordinates, without exploiting the intrinsic circumferential-longitudinal organization of the left ventricle (LV). We present MyoUV, a canonical UV surface- field framework for sparse SAX reconstruction. MyoUV maps epicardial and endocardial surfaces to anatomically oriented canonical UV domains, lifts sparse slice contours into dense UV conditioning through geometry-aware soft aggregation, and predicts continuous 3D coordinate fields with explicit circumferential periodicity. By replacing direct vertex-space regression with learning on a regular anatomical domain, MyoUV enables efficient convolutional reconstruction and a shared surface parameterization across subjects. Experiments on the multi-pathology ACDC dataset demonstrate improved geometric accuracy and inference efficiency under sparse SAX inputs.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/STACOM2026_032.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
Open Review Page: https://openreview.net/profile?id=~Yangang_Wang3 https://openreview.net/profile?id=~Lei_Li16
BibTex
@InProceedings{YuaXia_MyoUV_MICCAISAT2026,
author = { Yuan, Xiaohan AND Yang, Xuan AND Li, Qingya AND Lyu, Yilin AND Li, Lei AND Wang, Yangang},
title = { { MyoUV: Learning Canonical UV Surface Fields for Sparse SAX Myocardial Reconstruction } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17267},
month = {pending},
page = {pending}
}