New study compares loss functions for echocardiography segmentation with partial data

By PulseAugur Editorial · [2 sources] · 2026-07-06 12:50

A new study published on arXiv evaluates three loss functions—aCCE loss, marginal loss, and aBCE loss—for deep learning-based echocardiography segmentation using partially labeled data from multiple domains. The research found that all three functions perform well on intra-domain tasks. For inter-domain tasks, aBCE and marginal loss were superior when one label was missing, while marginal loss excelled when multiple labels were absent, demonstrating its robustness in complex scenarios. AI

IMPACT This research could lead to more robust AI models for medical image analysis, particularly in scenarios with incomplete or varied datasets.

RANK_REASON The cluster contains an academic paper detailing a comparison of loss functions for a specific deep learning task.

Read on arXiv cs.AI →

paper
other

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

New study compares loss functions for echocardiography segmentation with partial data

COVERAGE [2]

arXiv cs.AI TIER_1 English(EN) · Iman Islam, Esther Puyol-Ant\'on, Bram Ruijsink, Andrew J. Reader, Andrew P. King · 2026-07-07 04:00

Comparison of Loss Functions for Robust Deep Learning-based Echocardiography Segmentation when Learning with Partially Labelled Data from Multiple Domains

arXiv:2607.05008v1 Announce Type: cross Abstract: Echocardiography is the first imaging modality used for assessing cardiac function, and accurate segmentation of cardiac structures is essential for deriving biomarkers. However, the development of effective automated segmentation…
arXiv cs.AI TIER_1 English(EN) · Andrew P. King · 2026-07-06 12:50

Comparison of Loss Functions for Robust Deep Learning-based Echocardiography Segmentation when Learning with Partially Labelled Data from Multiple Domains

Echocardiography is the first imaging modality used for assessing cardiac function, and accurate segmentation of cardiac structures is essential for deriving biomarkers. However, the development of effective automated segmentation models for multiple cardiac structures is challen…

COVERAGE [2]

Comparison of Loss Functions for Robust Deep Learning-based Echocardiography Segmentation when Learning with Partially Labelled Data from Multiple Domains

Comparison of Loss Functions for Robust Deep Learning-based Echocardiography Segmentation when Learning with Partially Labelled Data from Multiple Domains

RELATED ENTITIES

RELATED TOPICS