162 docs tagged with "classification"

705 optic disc-centred fundus photographs (396 normal + 309 glaucoma) from Hospital Clinico San Carlos. Expert-annotated binary glaucoma classification.

1200 fundus images for AMD classification, five-class lesion segmentation, fovea/optic disc localisation.

One hundred retinal fundus images from Armed Forces Institute of Ophthalmology, Rawalpindi, with expert annotations for vessels and hypertensive retinopathy, diabetic retinopathy, and papilledema task

4800 AS-OCT images from 199 patients. Two tasks: angle closure classification and scleral spur localization.

~113,893 color fundus images labelled as referable glaucoma (RG), no referable glaucoma (NRG), or ungradable. Large-scale, multi-ethnic, multi-site screening dataset.

Two thousand preprocessed fundus images compiled from six public sources for normal, diabetes, cataract, and age-related macular degeneration classification.

De-identified fluorescein and indocyanine-green angiography images paired with structured lesion descriptions and reports.

Augmented ODIR-5K fundus photographs for 8-class ocular disease classification: Normal, Diabetes, Glaucoma, Cataract, AMD, Hypertension, Myopia, Other. Preprocessing includes CLAHE and standard augmen

3662 fundus images from Aravind Eye Hospital, graded 0-4 for diabetic retinopathy severity.

Retinal OCTA scans labeled for artifact type, artifact severity, signal strength, and image quality.

Surgical-microscope and intraoperative-OCT data for autonomous robotic retinal-vein cannulation in ex vivo porcine eyes.

Small four-class eye-disease image dataset with normal, cataract, glaucoma, and retinal disease categories.

One thousand seven hundred eighty-six original color fundus images for diabetic retinopathy, age-related macular degeneration, and glaucoma grading/classification.

Fundus photographs from Bangladesh Eye Hospital for glaucoma detection. Includes optic cup/disc crops and vessel segmentation masks alongside binary glaucoma/normal labels.

35,126 fundus images for 5-class diabetic retinopathy grading (No DR: 25,810 / Mild: 2,443 / Moderate: 5,292 / Severe: 873 / Proliferative: 708). Available via Kaggle.

Longitudinal widefield fundus photographs from birdshot chorioretinitis eyes and age- and sex-matched controls.

16,266 color fundus images from 8,524 Brazilian patients with 14 disease labels (DR, AMD, glaucoma, drusen, others), image quality flags, and demographic attributes.

Six near-infrared or synthetic iris subsets spanning close-range, lamp variation, twins, distance, large-scale, and synthetic recognition.

601 fundus images in 4 classes: Normal, Cataract, Glaucoma, and Retina Disease. Intended for ocular disease classification.

Extended phase/time labels for the Cataract-101 surgical-video dataset.

101 cataract surgery videos with 10-phase workflow annotations. Canonical older cataract benchmark.

3,000 cataract surgery procedures, 1,134.2 hours of video with phase annotations, instance segmentation, tracking, and skill scoring. Largest public cataract-surgery-video resource.

50 cataract surgery videos (>9 hours total) with frame-level annotations of 21 surgical tools. Earlier and larger sibling to Cataract-1K.

Fundus photographs and segmentation masks for subretinal fluid in central serous chorioretinopathy, with healthy-eye classification controls.

1,345 fundus images captured across multiple camera manufacturers with OD/OC segmentation + glaucoma classification labels.

Prospective tabular clinical dataset of 332 patients with chronic corneal disorders, including visual acuity, slit-lamp findings, comorbidities, diagnoses, treatments, and risk factors.

Composite fundus and OCT dataset with retinal layer, retinal lesion, and macular/glaucomatous disorder markings.

Swept-source OCT volumes for AMD and DME with three-dimensional pigment epithelial detachment and intraretinal-fluid masks.

Multicenter uveal-melanoma gene-expression, PRAME, clinical, and metastasis-free-survival data for prognostic modeling.

Professional/extended CORN database with in-vivo confocal microscopy images of the corneal subbasal nerve plexus. Supports nerve fiber segmentation, image quality enhancement, and tortuosity grading (

1,500 in-vivo confocal microscopy images of corneal subbasal nerve plexus, graded for tortuosity (4 levels). Train: 1,200 / Test: 300. 384×384 px, 400×400 µm FOV.

Combined corneal confocal microscopy collection comprising CORN-1, CORN-2, CORN-3, CORN-1500, CORN-Pro, and CORN-Complex. The subsets support corneal-nerve and cell segmentation, image enhancement, to

Fundus photographs grouped for corneal curvature prediction, with 50 images per group.

Tabular keratoconus dataset with 77 eye-level rows and 60 variables describing visual acuity, contrast sensitivity, stage, and corneal elevation/shape parameters.

Orbscan corneal topography maps and structured clinical annotations for keratoconus detection.

187 cataract cases with color fundus images and paired professional cataract-grading diagnostic reports. Designed for medical multimodal-LLM evaluation.

Fifty OCT and fundus images with healthy/glaucomatous labels and cup-to-disc ratio annotations by ophthalmologists.

12522 fundus images with DR grading (0-5) and lesion-level segmentation annotations.

2,000 regular fundus images (500 patients × 2 fields × 2 eyes) plus 256 ultra-widefield fundus images labeled for DR severity (ICDR grades 0-4) and image quality assessment (gradable/ungradable). From

15,709 fundus images with paired medical reports and extracted keywords. Only public fundus report-generation dataset — useful for VLM / captioning evaluation.

Fundus-image VQA dataset for diabetic macular edema derived from IDRiD and e-ophtha.

35,126 fundus images organised for 5-class DR severity grading (ICDR grades 0-4): No DR 25,810 / Mild 2,443 / Moderate 5,292 / Severe 873 / Proliferative 708. Hosted on Alibaba Tianchi.

174 OCTA images for DR lesion segmentation (IRMA, NPA, NV), image quality assessment, and DR grading (3-class).

Fifty color fundus images with expert markings of diabetic retinopathy lesions, blood vessels, optic disc, and macula.

3,100 paired two-field fundus images with DR grading labels. Only public two-field paired DR benchmark.

OCT data from 384 subjects, including 269 with AMD and 115 normal subjects, with 38,400 B-scans and derived total retina and RPE-drusen complex thickness measurements.

Forty-five retinal OCT volumes acquired with a Spectralis system: 15 dry AMD, 15 diabetic macular edema, and 15 normal volumes. The official Duke release provides the study data.

Precomputed image embeddings for BRSET and mBRSET to support efficient ophthalmic AI research without raw-image redistribution.

5,381 B-scan ocular ultrasound video clips with retinal-detachment presence and macula-on/off status labels. Total runtime approximately 5 hours 10 minutes. Only public ocular ultrasound video benchma

External eye photographs curated from publicly available web sources for binary classification of referable blepharitis versus normal.

Original and augmented eye-disease image dataset covering retinitis pigmentosa, retinal detachment, pterygium, myopia, macular scar, glaucoma, disc edema, diabetic retinopathy, central serous choriore

Raw and processed eye-tracking data from glaucoma patients with asymmetrical visual-field loss during free viewing.

Questionnaire and clinical eye-screening data from first-year university students in Ghana.

~102K VQA pairs derived from 58,485 images across 8 ophthalmic modalities (fluorescein angiography, ICGA, OCT, CFP, ultrasound biomicroscopy, slit-lamp, fundus auto-fluorescence, CT) covering 100+ dis

Visual-field data from glaucoma home monitoring: mean deviation, duration, and pointwise differential light sensitivity values across HFA and Eyecatcher tests.

~88,000 fundus images graded 0–4 for DR severity. Largest public DR dataset; competition images from EyePACS clinics.

Ophthalmic clinical text + NPZ records covering glaucoma, cataract, and neuro-ophthalmology with paired age, sex, race/ethnicity, and language attributes. Derived from the same Harvard clinical popula

Wide-field fundus images from premature infants for ROP staging (Stages 1–5 + Plus Disease). Expert-annotated for AI-assisted ROP diagnosis and screening.

Fundus photography, multispectral and functional retinal imaging, retinal blood-flow and pupillary-light videos, retina-characteristic labels, quality labels, demographics, and psychological assessmen

Repeated light/dark SD-OCT acquisitions and retinal reflectivity profiles from healthy, neuromyelitis-optica, and Alzheimer groups.

1,099 pediatric fundus images from 483 premature infants annotated for retinopathy of prematurity (ROP) staging. Standard fundus (RetCam) — not ultra-widefield; `rop_uwf_intelligent` is retained as th

1020 fundus images with binary glaucoma labels (697 normal, 323 glaucoma), plus optic disc/cup annotations.

300 paired fundus + 3D OCT volumes. Glaucoma grading into Normal / Early / Advanced plus OD/OC segmentation and fovea location.

Eye-level demographics, clinical measurements, and SD-OCT-derived macular ganglion-cell-complex thickness for glaucoma assessment.

300 circumpapillary OCT images. RNFL/GCIPL/choroid layer segmentation plus binary glaucoma classification.

263 eyes × 1,115 visits. Multi-modal: VF (HFA 24-2), fundus photographs, OCT RNFL, IOP, CCT. Labels: VF progression, OD segmentation, glaucoma stage.

1,000 patients with OCT RNFLT maps (225×225), visual field measurements, and demographics for glaucoma detection (binary) and longitudinal progression forecasting (6 definitions). First public glaucom

Fundus images for glaucoma detection from Harvard Medical School / Mass Eye and Ear. Binary glaucoma classification.

30,000 subjects (10K each AMD, DR, glaucoma) with paired SLO fundus and OCT B-scans, demographic attributes (race, ethnicity, gender, language), for fairness analysis.

HD-OCT scans and clinical outcome variables for macular-hole visual-improvement prediction after surgery.

A collection of 169 fundus photographs with expert exudate and bright lesion annotations, clinical metadata, optic nerve locations, vessel estimates, and image quality scores.

4,011 color fundus images labeled for high-myopia vs pathological-myopia classification — largest public dataset for HM/PM separation.

High-resolution fundus images from Hillel Yaffe Medical Center for age-related macular degeneration diagnosis.

747 fundus images from 288 patients with glaucoma labels confirmed by paired OCT + visual field. First public dataset with gold-standard multimodal glaucoma confirmation.

OCT challenge datasets from HDMILab covering retinal layer segmentation and fluid detection tasks from MICCAI/OMIA workshops. Includes sub-challenges such as AMD/CSC/DR classification and retinal laye

400 training fundus images with AMD classification (non-AMD vs AMD) and lesion annotations. Challenge dataset from ISBI 2020.

516 fundus images with DR grade (0-4), macular edema grade (0-2), and pixel-level lesion segmentation for 81 images.

Image and symptom dataset for normal, uveitis, conjunctivitis, cataract, and eyelid-drooping classification, with source-reported SMOTE balancing to 649 images per class.

Retinal fundus dataset for five-class diabetic-retinopathy grading.

9,939 color fundus images from Japanese patients with DR grading labels (Davis grading scale).

1,000 fundus images spanning 39 ophthalmic disease categories from the Joint Shantou International Eye Center. Used for multi-class fundus disease classification.

101,442 gradable fundus images labeled as referable (RG) or non-referable (NRG) for glaucoma. Image data from EyePACS LLC; labels from Rotterdam Eye Hospital expert graders.

Small corneal confocal fluorescence imaging dataset for keratoconus epithelium analysis.

Corneal map images for three-class keratoconus detection.

~84,000 retinal OCT B-scan images across 4 classes: CNV, DME, DRUSEN, NORMAL. Train: ~83,484 / Test: 1000.

11,760 fundus images with glaucoma classification labels and ophthalmologist-derived attention maps. Largest public glaucoma dataset with attention ground truth.

Text/tabular dataset supporting readability and language-access analyses in ophthalmology.

Light-adapted electroretinogram and oscillatory-potential waveform dataset from 253 pediatric participants, with structured JSON, tabular metadata, and electrode-position eye images.

Composite multimodal ophthalmology benchmark with multi-granular anatomical, diagnostic, staging, demographic, and text annotations.

Anonymized retinal fundus images from an ophthalmology clinic in M'Sila, Algeria for diabetic-retinopathy classification.

198 fundus images (from MESSIDOR) with pixel-wise segmentation for 10 biomarkers (lesions + structures) + DR and ME severity grades.

~30,000 longitudinal OCT B-scans across multiple patient visits, annotated for AMD change detection and progression monitoring.

5,164 fundus images from 1,291 patients captured with the Phelcom Eyer handheld smartphone-based fundus camera. Labels for DR grading + clinical/demographic prediction.

6,272 AS-OCT images + 392 anterior-segment photographs for corneal opacity assessment with expert grading. First large public multimodal AS-OCT + photo dataset for cornea.

1748 fundus images with DR grading (0-3 Retinopathy Grade) and macular edema risk (0-2).

34,452 fundus images from 4,344 patients across multiple fields per eye, with DR screening labels. Largest public four-field DR dataset.

Ophthalmology-specific multimodal reasoning dataset built from 45 public datasets. Chain-of-thought reasoning traces for retinal VQA.

Multi-modal DR dataset combining color fundus photographs (CFP), OCT B-scans, and ultra-widefield (UWF) fundus images, annotated for DR detection and severity grading. Published in Nature Scientific D

Pentacam-derived corneal tomography variables for intraocular-lens and corneal subtype modeling.

Multi-source heterogeneous retinal fundus image-quality assessment dataset for training and evaluating quality-control models.

A multicenter collection of 186 minimally invasive glaucoma surgery (MIGS) videos with annotations for surgical-phase recognition and semantic segmentation of instruments and anatomical structures.

58,036 fundus + 45,923 OCT images assembled for multi-disease classification (8 classes) with cross-modal distillation. Sourced from multiple public ophthalmic datasets.

2,208 color fundus images with 20-class multi-label disease annotations.

Public sample fundus photographs and measurements for modeling myopic regression after corneal refractive surgery.

Longitudinal near-infrared iris images with subject, eye, age, sex, and ethnicity metadata used in ICE iris-recognition evaluations.

Retinal OCT B-scans from Noor Eye Hospital for classification of Normal, Drusen, and CNV (choroidal neovascularisation) cases. 16,822 B-scans from 441 eyes (Normal 120 / Drusen 160 / CNV 161 eyes).

Linked color fundus and macular OCT images for diabetic macular edema and diabetic retinopathy classification, with CSV labels and shared patient/eye/image nomenclature.

A set of 1,110 optic nerve head OCT volumes from 624 patients, including 847 scans with primary open angle glaucoma and 263 healthy scans.

1,672 OCT B-scans from AMD, DME, and healthy controls with 5,016 multi-grader layer annotations + lesion detection labels.

A derived OCT angiography resource with 640 coronal PNG views for each of 129 subjects: 90 normal, 29 diabetic retinopathy, 5 AMD, and 5 choroidal neovascularization cases.

500 subjects with OCTA volumes, vessel segmentation, FAZ (foveal avascular zone) annotations, and layer segmentation. Largest public OCTA dataset.

2,000+ OCT images labeled for 7 conditions: AMD, DME, ERM, NO (normal), RAO, RVO, VID.

500 OCT images: NORMAL (206), AMD (50), CSC (128), DR (59), MH (57). High-resolution B-scans for 5-class classification.

844,000 simulated patient-physician dialogue rows generated from AREDS clinical visits. Enables ophthalmic dialogue and counseling VLM training.

16,530 ultra-widefield fundus images from 8,405+ patients (age 0–90) annotated for 38 ophthalmic diseases and 67 fundus features. Released alongside the FairerOPTH study on sexism and ageism in ophtha

8000 patients (16000 fundus images, left + right eye) with 8 disease labels for multi-label ocular disease classification.

Longitudinal OCT + fundus dataset from AMD/DME patients across multiple clinical visits. 9,408 OCT B-scans have biomarker labels for 8 categories (fluid, drusen, scarring, PED variants, etc.); 78,000+

700 Optos ultra-wide-field fundus images with multi-disease classification labels and image-quality flags from real-world clinical practice.

Large-scale multi-procedure ophthalmic surgical video dataset covering 66 surgery types, 102 phases, 150 operations (~285 h). 1,969 untrimmed videos; 17,508 trimmed operation-level clips; 14,674 trimm

160,185 video clip-instruction pair samples from 9,819 ophthalmic surgical videos, covering multiple procedure types. Designed for text-guided surgical video generation and understanding. Published at

Ophthalmic visual question-answering benchmark dataset released as supplementary data for evaluating multimodal language models in ophthalmology.

Paired tabletop and portable retinal images from the same patients, enabling cross-device domain-adaptation research.

1200 fundus images for pathologic myopia classification and optic disc / lesion segmentation. 50 % PM / 50 % non-PM.

488 fundus images from 244 patients (both eyes). Expert OD/OC segmentation, glaucoma stage, IOP, and clinical metadata.

757 color fundus images from a Paraguayan cohort with 7-class DR grading labels. Adds Latin-American representation.

Transient pattern electroretinogram responses from 304 subjects in 336 records with clinical metadata.

Fourteen thousand nine hundred seventy-six near-infrared ocular images for eye-state detection and coarse gaze-direction classification during ophthalmic measurement workflows.

Fundus-photography dataset for retinal artery occlusion diagnosis, based on web-derived public data and public fundus datasets.

148 Heidelberg Spectralis SD-OCT volumes (~4,254 B-scans) for 3-class volume-level classification: Normal (50 volumes), AMD (48 volumes), DME (50 volumes). Variable B-scans per volume: 19, 25, 31, or

1200 fundus images: 400 train, 400 val, 400 test. Labels: glaucoma/non-glaucoma + optic disc/cup segmentation.

1200 fundus images from REFUGE1 with optic disc/cup segmentation masks annotated by multiple expert raters.

2000 fundus images for glaucoma classification, optic disc/cup segmentation, and fovea localisation. Extends REFUGE1.

Baseline and two-year follow-up color fundus image pairs from Tianjin Medical University for diabetic-retinopathy progression research.

6,004 pediatric RetCam fundus images from 188 newborns in Ostrava (Czech Republic), annotated for retinopathy of prematurity (ROP) screening + classification.

~24,000 retinal OCT images across 8 disease classes: AMD, BRAO, BRVO, CSC, CRAO, CRVO, DME, MH.

Auxiliary RFMiD 2.0 retinal fundus dataset spanning 51 disease categories.

3200 fundus images annotated for 45 retinal conditions. Used for multi-label disease classification.

485 fundus images (313 normal, 172 glaucoma) with optic disc region of interest crops for deep learning.

Clinical records combining retinal nerve fiber layer, visual-field, corneal-thickness, and intraocular-pressure features.

100 fundus images with microaneurysm annotations from the Retinopathy Online Challenge.

Grand Challenge OCT classification dataset for diabetic-retinopathy related OCT classification.

3,012 OCT B-scans from 146 eyes / 130 patients with retinal vein occlusion. Dual-task labels: fluid segmentation + retinal-layer segmentation. Detection of macular lesions.

Small-incision cataract surgery video dataset with train/validation/test video archives for phase-recognition research.

~2,617 annotated slit-lamp frames covering anterior-segment anatomy + multi-lesion detection (cataract, corneal disease, conjunctivitis).

~12,000 fundus images aggregated from 19 public glaucoma datasets with standardized disc/cup/vessel channels and unified labels. CC0 — fully public domain.

Longitudinal OCT angiography projection maps with vessel labels, clinical text, and treatment/follow-up groupings.

1,219 color fundus images with DR severity grading + pixel-level exudate segmentation masks. Multi-center (SUSTech + Sun Yat-sen University).

Synthetic circumpapillary OCT images for healthy and glaucomatous eyes with retinal-layer masks and RNFL thickness values.

Retinopathy-of-prematurity fundus-image dataset and synthetic-image resources for privacy-preserving AI training.

Head-mounted eye images with pupil, iris, eyelid, eyeball, gaze, landmark, segmentation, and eye-movement annotations.

A collection of 1,800 preprocessed retinal OCT B-scans, with 600 images each for AMD, diabetic macular edema, and normal retina.

External ocular photographs of thyroid eye disease collected from web sources, with surgery-related images removed by the source curator.

Fundus photographs for ocular toxoplasmosis detection: active toxoplasmosis, inactive (scarred) lesions, and normal. ~412 images total (adult + pediatric cases).

13,047 ultra-widefield fundus photographs (Optos, 200° FOV) from Tsukazaki Hospital, annotated with 8 binary disease labels: AO, AMD, DR, Glaucoma, MH, RD, RP, RVO. Associated with Scientific Data 202

Visible-light iris images captured at a distance and on the move with realistic blur, reflection, occlusion, pose, and illumination noise.

Ultra-widefield fundus photographs for diabetic retinopathy grading (5-level ICDR scale). Released alongside the Reasoning-Enhanced VLM paper for interpretable UWF DR detection. Related to the UWF4DR

1,630 Optos ultra-wide-field fundus images from 809 patients graded for diabetic retinopathy (5-class ICDR) by senior ophthalmologists.

2,031 ultra-widefield fundus images for AI-assisted intraocular tumor detection and classification. Six categories: Normal, Choroidal Hemangioma (CH), Retinal Capillary Hemangioma (RCH), Choroidal Ost

28,943 HFA 24-2 visual field tests with per-point sensitivities, MD/PSD indices, glaucoma labels, and longitudinal follow-up data. Input: 52-element sensitivity array (not an image).

A multilabel fundus classification challenge with 3,435 labeled training images and 350 test images covering six disease groups and normal findings.

Visual-field and psychophysics experiment data for scotoma-detection comparisons.

Ophthalmic image-text VQA/reasoning benchmark with retinal modalities and expert-verified question-answer pairs.