<b>MAPLES-DR: </b>MESSIDOR Anatomical and Pathological Labels for Explainable Screening of Diabetic Retinopathy
Notice bibliographique
Résumé
<b>MAPLES-DR</b> (MESSIDOR Anatomical and Pathological Labels for Explainable Screening of Diabetic Retinopathy) [1], contains <i>new diagnoses</i> for Diabetic Retinopathy (DR) and Macular Edema (ME) as well as <i>new pixel-wise </i><i>segmentation maps</i> for 10 retinal structures related to those pathologies for 198 images of the MESSIDOR public fundus dataset [2].<br>The annotation procedure and an evaluation of the MAPLES-DR labels are documented in a Scientific Data paper [1]. If you wish to use this dataset in academic works, we kindly ask you to cite:<br>Gabriel Lepetit-Aimon, Clément Playout, Marie Carole Boucher, Renaud Duval, Michael H Brent, and Farida Cheriet. Maples-dr: messidor anatomical and pathological labels for explainable screening of diabetic retinopathy. <i>Scientific Data</i>, 11(1):914, 08 2024. doi:10.1038/s41597-024-03739-6.<br>To ease the integration of MAPLES-DR with MESSIDOR images, we published a Python package named maples_dr that automates the download of MAPLES-DR labels, and the matching, cropping and resizing of MESSIDOR fundus images. We strongly encourage AI researchers willing to train models on MAPLES-DR labels to visit our Quick Start page.<b>Description of the Dataset</b>Segmentation of Retinal StructuresFrom February 2019 to February 2020, seven senior retinologists from Montréal and Toronto (Canada) segmented anatomical and pathological structures related to DR on 198 fundus images extracted from the public dataset MESSIDOR [2]:4 Anatomical Structures: Optic Disc, Optic Cup, Macula, and Vessels;3 Bright Lesions: Exudates, Cotton Wool Spots, and Drusens;3 Red Lesions: Microaneurysms, Hemorrhages, and Neovessels.These 10 biomarkers are provided as individual PNG binary images, along with two other segmentation maps: <b>Uncertain Bright</b><i> </i>and <b>Uncertain Red,</b> which contain structures that are clearly pathological but whose exact nature is unclear (e.g. microaneurysm vs. hemorrhage for a red lesion).For more details on the biomarkers definition or on their role in the diagnosis of DR, please refer to the MAPLES-DR Labels page.DR and ME gradesMAPLES-DR grades for DR and ME follow the guidelines developed for Canadian teleopthalmology screening. These guidelines distinguish seven grades for DR (<b>R0</b>: absent, <b>R1</b>: mild, <b>R2</b>: moderate, <b>R3</b>: severe, <b>R4A</b>: proliferative, <b>R4S</b>: stable treated proliferative, <b>R6</b>: insufficient image quality) and four for ME (<b>M0</b>: absent, <b>M1</b>: mild, <b>M2</b>: moderate, <b>M6</b>: insufficient image quality). These grades are defined systematically by the number and position of visible red and bright retinal lesions. Each grade is associated with a recommended course of action (from rescreening in two years for mild cases, to immediate referral to an ophthalmologist for the more severe ones). A detailed definition of the grading system can be found in this paper [3].<b>Data Records</b>MAPLES-DR dataset is distributed as two archives: <b>MAPLES-DR.zip</b> and <b>AdditionalData.zip</b>. The first one contains the main data of MAPLES-DR (segmentation maps and grades), while the second one contains additional information on the annotation process (time, comments) as well as intermediate data (pre-annotation maps, grades before consensus...).Please consult MAPLES-DR Data Records for more information.The additional archive <b>LesionVariabilityStudy.zip</b> contains the segmentation maps collected for a study measuring inter-observer variability when manually segmenting micro-aneurysms, haemorrhages, hard exudates and cotton-wool spots.<b>Usage</b>The annotations provided in MAPLES-DR are published under the CC-BY license. If you wish to use this dataset in academic work, we kindly ask you to cite the following paper [1]:<br>Gabriel Lepetit-Aimon, Clément Playout, Marie Carole Boucher, Renaud Duval, Michael H Brent, and Farida Cheriet. Maples-dr: messidor anatomical and pathological labels for explainable screening of diabetic retinopathy. <i>Scientific Data</i>, 11(1):914, 08 2024. doi:10.1038/s41597-024-03739-6.<br>Note that the fundus images corresponding to the diagnostic labels and biomarker segmentation maps of MAPLES-DR are the property of the MESSIDOR consortium and are freely available from the Consortium's website. For more specific instructions see Using MESSIDOR images.To ease the integration of the two datasets we published a Python package named maples_dr that automates MAPLES-DR download, matching with MESSIDOR images, cropping and resizing to a uniform image resolution. We strongly encourage AI researchers willing to train models on MAPLES-DR labels to visit our Quick Start page.AcknowledgmentsThis study was funded by the NSERC (Natural Science and Engineering Research Council of Canada) as well as Diabetes Action Canada and FROUM (Fonds de recherche en ophtalmologie de l’Université de Montréal).The original MESSIDOR dataset was kindly provided by the Messidor program partners (see https://www.adcis.net/en/third-party/messidor/).References[1] Gabriel Lepetit-Aimon, Clément Playout, Marie Carole Boucher, Renaud Duval, Michael H Brent, and Farida Cheriet. Maples-dr: messidor anatomical and pathological labels for explainable screening of diabetic retinopathy. <i>Scientific Data</i>, 11(1):914, 08 2024. doi:10.1038/s41597-024-03739-6.[2] Etienne Decencière, Xiwei Zhang, Guy Cazuguel, Bruno Lay, Béatrice Cochener, Caroline Trone, Philippe Gain, Richard Ordonez, Pascale Massin, Ali Erginay, and others. Feedback on a publicly distributed image database: the MESSIDOR database. <i>Image Analysis & Stereology</i>, 33(3):231–234, Aug 2014. doi:http://dx.doi.org/10.5566/ias.1155.[3] M.C. Boucher, J. Qian, M.H. Brent, D.T. Wong, T. Sheidow, R. Duval, A. Kherani, R. Dookeran, D. Maberley, A. Samad, and V. Chaudhary. Evidence-based Canadian guidelines for tele-retina screening for diabetic retinopathy: recommendations from the Canadian Retina Research Network (CR2N) Tele-Retina Steering Committee. <i>Canadian Journal of Ophthalmology</i>, 55(1):14–24, Feb 2020. doi:10.1016/j.jcjo.2020.01.001.<br>
Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.
Comment cette classification a été obtenuedéplier
Prédiction distillée sur la base complète
Imitation des enseignantsNi prévalence calibrée, ni vérité terrain. Validation humaine à venir. Apprise à partir de 10 348 étiquettes directes de Codex et de 10 348 étiquettes directes de Gemma. Le mode candidate est l'union des têtes enseignantes seuillées; le consensus est leur intersection. Ces sorties portent le statut machine_predicted_unvalidated et ne sont ni des étiquettes humaines ni des étiquettes directes de modèles de pointe.
Scores Codex et Gemma par catégorie
| Catégorie | Codex | Gemma |
|---|---|---|
| Métarecherche | 0,001 | 0,001 |
| Méta-épidémiologie (sens strict) | 0,000 | 0,000 |
| Méta-épidémiologie (sens large) | 0,001 | 0,000 |
| Bibliométrie | 0,000 | 0,000 |
| Études des sciences et des technologies | 0,000 | 0,000 |
| Communication savante | 0,000 | 0,000 |
| Science ouverte | 0,000 | 0,000 |
| Intégrité de la recherche | 0,000 | 0,000 |
| Charge utile insuffisante (le modèle a refusé de juger) | 0,000 | 0,000 |
Scores machine (provisoires)
Les deux têtes enseignantes du modèle étudiant, lues sur ce travail. Un score ordonne la base pour la relecture; il n'affirme jamais une catégorie, et le statut de validation accompagne chaque rangée tel quel.
Scores de référence d'un modèle non mature (critères de maturité non atteints, 7 itérations). Un score ordonne; il n'affirme jamais une catégorie.
score_only:v0-immature-baseline · tel quel depuis la passe de notation : score_only signifie que le nombre peut ordonner les travaux, et qu'aucune étiquette de catégorie n'en découleClassification
machine, non validéePrédiction automatique; un appel candidat d’une seule tête enseignante, pas un consensus.
Le détail, modèle par modèle et score par score, se trouve en fin de page sous « Comment cette classification a été obtenue ».