Deep Learning Segmentation in 2D echocardiography using the CAMUS dataset : Automatic Assessment of the Anatomical Shape Validity

TL;DR

This study extends deep learning evaluation for 2D echocardiography segmentation by incorporating anatomical correctness, providing new insights into model ranking beyond traditional metrics.

Contribution

It introduces an evaluation framework that assesses the anatomical validity of segmentation shapes, enhancing the assessment of deep learning models for cardiac ultrasound.

Findings

Anatomical assessment alters model ranking

Traditional metrics may not reflect shape correctness

Proposed criteria improve clinical relevance

Abstract

We recently published a deep learning study on the potential of encoder-decoder networks for the segmentation of the 2D CAMUS ultrasound dataset. We propose in this abstract an extension of the evaluation criteria to anatomical assessment, as traditional geometric and clinical metrics in cardiac segmentation do not take into account the anatomical correctness of the predicted shapes. The completed study sheds a new light on the ranking of models.