Standardized evaluation system for left ventricular segmentation algorithms in 3D echocardiography
Olivier Bernard
(1, 2)
,
J.G.H. Bosch
(3)
,
B Heyde
(4)
,
M. Alessandrini
(4)
,
D. Barbosa
(5)
,
S. Camarasu-Pop
(6, 1)
,
F. Cervenansky
(6)
,
Sébastien Valette
(7)
,
O. Mirea
(4)
,
M. Bernier
(8)
,
P.-M. Jodoin
(8)
,
J.-S. Domingos
(9)
,
R.V Stebbing
(9)
,
K Keraudren
(10)
,
O. Oktay
(10)
,
J. Caballero
(10)
,
W. Shi
(10)
,
D. Rueckert
(10)
,
F. Milletari
(11)
,
F. Ahmadi
(12)
,
E. Smistad
(13)
,
F. Lindseth
(13)
,
M van Stralen
(14)
,
Chen Wang
(15, 16)
,
O. Smedby
(16, 15)
,
Erwan Donal
(17)
,
M. Monaghan
(18)
,
A. Papachristidis
(18)
,
M.L. Geleijnse
(3)
,
Elena Galli
(17)
,
J. d'Hooge
(4)
1
Images et Modèles
2 MYRIAD - Modeling & analysis for medical imaging and Diagnosis
3 Thoraxcenter Biomedical Engineering
4 Laboratory of Cardiovascular Imaging and Dynamics
5 ICVS - Life and Health Sciences Research Institute [Braga]
6 Service Informatique et développements
7 CREATIS - Centre de Recherche en Acquisition et Traitement de l'Image pour la Santé
8 UdeS - Université de Sherbrooke
9 IBME - Institute of Biomedical Engineering
10 Department of Computing [London]
11 CAMPAR - Computer Aided Medical Procedures & Augmented Reality
12 Department of Neurology
13 NTNU - Norwegian University of Science and Technology [Trondheim]
14 ISI - Image sciences institute - University of Utrecht
15 School of Technology and Health
16 Department of Medical and Health Sciences (IMH)
17 LTSI - Laboratoire Traitement du Signal et de l'Image
18 KCH - King's College Hospital
2 MYRIAD - Modeling & analysis for medical imaging and Diagnosis
3 Thoraxcenter Biomedical Engineering
4 Laboratory of Cardiovascular Imaging and Dynamics
5 ICVS - Life and Health Sciences Research Institute [Braga]
6 Service Informatique et développements
7 CREATIS - Centre de Recherche en Acquisition et Traitement de l'Image pour la Santé
8 UdeS - Université de Sherbrooke
9 IBME - Institute of Biomedical Engineering
10 Department of Computing [London]
11 CAMPAR - Computer Aided Medical Procedures & Augmented Reality
12 Department of Neurology
13 NTNU - Norwegian University of Science and Technology [Trondheim]
14 ISI - Image sciences institute - University of Utrecht
15 School of Technology and Health
16 Department of Medical and Health Sciences (IMH)
17 LTSI - Laboratoire Traitement du Signal et de l'Image
18 KCH - King's College Hospital
Olivier Bernard
- Fonction : Auteur
- PersonId : 172471
- IdHAL : bernard-creatis
- ORCID : 0000-0003-0752-9946
M. Alessandrini
- Fonction : Auteur
- PersonId : 944183
S. Camarasu-Pop
- Fonction : Auteur
- PersonId : 19719
- IdHAL : camarasu
- ORCID : 0000-0002-7923-5069
- IdRef : 178745596
F. Cervenansky
- Fonction : Auteur
- PersonId : 19446
- IdHAL : frederic-cervenansky
- IdRef : 278760546
Sébastien Valette
- Fonction : Auteur
- PersonId : 20259
- IdHAL : sebastien-valette
- ORCID : 0000-0001-7549-4808
- IdRef : 070323496
M. Bernier
- Fonction : Auteur
- PersonId : 843320
- ORCID : 0000-0002-7812-4965
- IdRef : 116459301
Chen Wang
- Fonction : Auteur
- PersonId : 829307
Erwan Donal
- Fonction : Auteur
- PersonId : 947765
- ORCID : 0000-0003-2677-3389
Résumé
Real-time 3D Echocardiography (RT3DE) has been proven to be an accurate tool for left ventricular (LV) volume assessment. However, identification of the LV endocardium remains a challenging task, mainly because of the low tissue/blood contrast of the images combined with typical artifacts. Several semi and fully automatic algorithms have been proposed for segmenting the endocardium in RT3DE data in order to extract relevant clinical indices, but a systematic and fair comparison between such methods has so far been impossible due to the lack of a publicly available common database. Here, we introduce a standardized evaluation framework to reliably evaluate and compare the performance of the algorithms developed to segment the LV border in RT3DE. A database consisting of 45 multivendor cardiac ultrasound recordings acquired at different centers with corresponding reference measurements from 3 experts are made available. The algorithms from nine research groups were quantitatively evaluated and compared using the proposed online platform. The results showed that the best methods produce promising results with respect to the experts’ measurements for the extraction of clinical indices, and that they offer good segmentation precision in terms of mean distance error in the context of the experts’ variability range. The platform remains open for new submissions.