An explainable ensemble for diabetic retinopathy grading with a novel confidence quality factor and configurable heatmaps

While current artificial intelligence (AI) tools aid in detecting diabetic retinopathy (DR), they face significant challenge that limit their clinical utility. Most are restricted to binary (referable vs. non-referable) screening and operate as “black boxes,” lacking the detailed, transparent explan...

Full description

Bibliographic Details
Authors: Civit Masot, Javier, Luna Perejón, Francisco, Muñoz Saavedra, Luis, Rodríguez Corral, José María, Domínguez Morales, Manuel Jesús, Civit Breu, Antón
Format: article
Status:Published version
Publication Date:2026
Country:España
Institution:Universidad de Sevilla (US)
Repository:idUS. Depósito de Investigación de la Universidad de Sevilla
OAI Identifier:oai:idus.us.es:11441/183026
Online Access:https://hdl.handle.net/11441/183026
https://doi.org/10.1007/s11517-026-03514-2
Access Level:Open access
Keyword:Diabetic retinopathy
Explainable AI (xAI)
Deep learning
Explainable ensemble
Medical imaging
ICDR
Clinical decision support
id ES_eaabdacc0e8cad25a4e626273b69b3ee
oai_identifier_str oai:idus.us.es:11441/183026
network_acronym_str ES
network_name_str España
repository_id_str
spelling An explainable ensemble for diabetic retinopathy grading with a novel confidence quality factor and configurable heatmapsCivit Masot, JavierLuna Perejón, FranciscoMuñoz Saavedra, LuisRodríguez Corral, José MaríaDomínguez Morales, Manuel JesúsCivit Breu, AntónDiabetic retinopathyExplainable AI (xAI)Deep learningExplainable ensembleMedical imagingICDRClinical decision supportWhile current artificial intelligence (AI) tools aid in detecting diabetic retinopathy (DR), they face significant challenge that limit their clinical utility. Most are restricted to binary (referable vs. non-referable) screening and operate as “black boxes,” lacking the detailed, transparent explanations required for diagnostic confidence. This study addresses these gaps by introducing a novel, explainable ensemble-based approach for detailed DR grading. Our system utilizes a parallel ensemble of two efficient deep learning networks, EfficientNetV2 and ConvNeXt, to perform a full five-class international clinical diabetic retinopathy (ICDR) classification. The proposed model achieves state-of-the-art performance, with 96.7% accuracy and an Area Under the Curve (AUC) over 96% for all classes on a public dataset. More importantly, it provides a comprehensive diagnostic report designed to enhance clinical trust and utility. This report features multiple, configurable superimposed heatmaps, two probability-ordered diagnostic suggestions, and a novel quality factor that estimates the confidence of the prediction. By offering richer, more transparent, and interactive explanations, our system moves beyond simple screening to function as a valuable diagnostic assistance tool for ophthalmologists and other healthcare professionals.SpringerArquitectura y Tecnología de ComputadoresTEP108: Robótica y Tecnología de ComputadoresEuropean Union (UE)2026info:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersionapplication/pdfapplication/pdfhttps://hdl.handle.net/11441/183026https://doi.org/10.1007/s11517-026-03514-2reponame:idUS. Depósito de Investigación de la Universidad de Sevillainstname:Universidad de Sevilla (US)InglésMedical and Biological Engineering and Computing.TSI-100930-2023-2https://link.springer.com/article/10.1007/s11517-026-03514-2info:eu-repo/semantics/openAccessoai:idus.us.es:11441/1830262026-06-17T12:51:07Z
dc.title.none.fl_str_mv An explainable ensemble for diabetic retinopathy grading with a novel confidence quality factor and configurable heatmaps
title An explainable ensemble for diabetic retinopathy grading with a novel confidence quality factor and configurable heatmaps
spellingShingle An explainable ensemble for diabetic retinopathy grading with a novel confidence quality factor and configurable heatmaps
Civit Masot, Javier
Diabetic retinopathy
Explainable AI (xAI)
Deep learning
Explainable ensemble
Medical imaging
ICDR
Clinical decision support
title_short An explainable ensemble for diabetic retinopathy grading with a novel confidence quality factor and configurable heatmaps
title_full An explainable ensemble for diabetic retinopathy grading with a novel confidence quality factor and configurable heatmaps
title_fullStr An explainable ensemble for diabetic retinopathy grading with a novel confidence quality factor and configurable heatmaps
title_full_unstemmed An explainable ensemble for diabetic retinopathy grading with a novel confidence quality factor and configurable heatmaps
title_sort An explainable ensemble for diabetic retinopathy grading with a novel confidence quality factor and configurable heatmaps
dc.creator.none.fl_str_mv Civit Masot, Javier
Luna Perejón, Francisco
Muñoz Saavedra, Luis
Rodríguez Corral, José María
Domínguez Morales, Manuel Jesús
Civit Breu, Antón
author Civit Masot, Javier
author_facet Civit Masot, Javier
Luna Perejón, Francisco
Muñoz Saavedra, Luis
Rodríguez Corral, José María
Domínguez Morales, Manuel Jesús
Civit Breu, Antón
author_role author
author2 Luna Perejón, Francisco
Muñoz Saavedra, Luis
Rodríguez Corral, José María
Domínguez Morales, Manuel Jesús
Civit Breu, Antón
author2_role author
author
author
author
author
dc.contributor.none.fl_str_mv Arquitectura y Tecnología de Computadores
TEP108: Robótica y Tecnología de Computadores
European Union (UE)
dc.subject.none.fl_str_mv Diabetic retinopathy
Explainable AI (xAI)
Deep learning
Explainable ensemble
Medical imaging
ICDR
Clinical decision support
topic Diabetic retinopathy
Explainable AI (xAI)
Deep learning
Explainable ensemble
Medical imaging
ICDR
Clinical decision support
description While current artificial intelligence (AI) tools aid in detecting diabetic retinopathy (DR), they face significant challenge that limit their clinical utility. Most are restricted to binary (referable vs. non-referable) screening and operate as “black boxes,” lacking the detailed, transparent explanations required for diagnostic confidence. This study addresses these gaps by introducing a novel, explainable ensemble-based approach for detailed DR grading. Our system utilizes a parallel ensemble of two efficient deep learning networks, EfficientNetV2 and ConvNeXt, to perform a full five-class international clinical diabetic retinopathy (ICDR) classification. The proposed model achieves state-of-the-art performance, with 96.7% accuracy and an Area Under the Curve (AUC) over 96% for all classes on a public dataset. More importantly, it provides a comprehensive diagnostic report designed to enhance clinical trust and utility. This report features multiple, configurable superimposed heatmaps, two probability-ordered diagnostic suggestions, and a novel quality factor that estimates the confidence of the prediction. By offering richer, more transparent, and interactive explanations, our system moves beyond simple screening to function as a valuable diagnostic assistance tool for ophthalmologists and other healthcare professionals.
publishDate 2026
dc.date.none.fl_str_mv 2026
dc.type.none.fl_str_mv info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
format article
status_str publishedVersion
dc.identifier.none.fl_str_mv https://hdl.handle.net/11441/183026
https://doi.org/10.1007/s11517-026-03514-2
url https://hdl.handle.net/11441/183026
https://doi.org/10.1007/s11517-026-03514-2
dc.language.none.fl_str_mv Inglés
language_invalid_str_mv Inglés
dc.relation.none.fl_str_mv Medical and Biological Engineering and Computing.
TSI-100930-2023-2
https://link.springer.com/article/10.1007/s11517-026-03514-2
dc.rights.none.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
application/pdf
dc.publisher.none.fl_str_mv Springer
publisher.none.fl_str_mv Springer
dc.source.none.fl_str_mv reponame:idUS. Depósito de Investigación de la Universidad de Sevilla
instname:Universidad de Sevilla (US)
instname_str Universidad de Sevilla (US)
reponame_str idUS. Depósito de Investigación de la Universidad de Sevilla
collection idUS. Depósito de Investigación de la Universidad de Sevilla
repository.name.fl_str_mv
repository.mail.fl_str_mv
_version_ 1869423160160419840
score 15,812455

Similar Items