Abstract
Automated echocardiographic segmentation remains challenging in low-resource clinical environments due to limited expert annotations and substantial variation between imaging views. Although multiple complementary views are routinely acquired during cardiac examinations, existing approaches typically process each view independently and overlook the anatomical correspondence available between acquisitions from the same patient.
In this work, we propose Patient-Paired Anatomical Consistency Learning, a framework that exploits naturally paired multi-view echocardiographic acquisitions as an additional source of supervision for cross-view segmentation. Unlike conventional domain adaptation approaches that align independent source and target distributions, our method leverages patient-level correspondence to learn anatomically consistent cardiac representations. We introduce anatomical representation consistency and functional consistency objectives that encourage complementary views of the same patient to share clinically meaningful features while preserving spatial information required for accurate segmentation.
We evaluate our framework on the CAMUS dataset under the challenging 4CH→2CH cross-view generalization setting, where target-view annotations are unavailable during training. Our results demonstrate that patient-paired anatomical supervision improves segmentation robustness and left ventricular functional measurement consistency compared with conventional feature alignment strategies.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/MIRASOL_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/forum?id=mUJvAGSyfI
BibTex
@InProceedings{AboIle_PatientPaired_MICCAISAT2026,
author = { Abolade, Ilerioluwakiiye},
title = { { Patient-Paired Anatomical Consistency Learning for Cross-View Echocardiographic Segmentation } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17264},
month = {pending},
page = {pending}
}