Core Areas of Focus

Diabetic Retinopathy
& Macular Disease
Non-dilated widefield fundus and macular OCT structured for DR severity grading, DME detection, fluid/thickness endpoints, and longitudinal progression analysis.
Glaucoma Risk
OCT RNFL/GCC and related structural markers organized for glaucoma-risk stratification, thinning-pattern analysis, and clinician-linked verification workflows.
Oculomics & Cardiometabolic Risk
Widefield fundus, OCT, and cardiometabolic anchors linked to support blood pressure, cardiovascular, and metabolic risk inference with grounded clinical context.
Video Capture &
External-Eye Imaging
Smartphone-based 15/30-second video plus still capture, standardized for cross-device acquisition, color consistency, and pseudonymous linkage to fundus/OCT.

Available as versioned Release
Packs with expert-adjudicated labels, governed metadata,
and audit-ready documentation.

We bridge the structural divide between raw clinical records & regulatory-ready AI training cohorts through controlled lineage, validation checkpoints, and documentation packaging.

Provenance Mapping

End-to-end traceability across every
transformation stage.

IRR-Verified Annotation

Multi-Clinician review with adjudication and
inter-rater reliability controls.

Compliance Packaging

Audit-ready documentation structured for
regulated AI diligence and review.

Engineered for Regulatory Defensibility

Engineered for
Regulatory Defensibility

Native integration of EU AI Act Article 10 aligned technical documentation. Every cohort is delivered with comprehensive data provenance, bias detection metrics, and a regulatory audit trail.

Beyond simple de-identification. We employ ISO 27001-aligned controls and HIPAA/GDPR-grade privacy layers, ensuring data purity remains intact from the point of care to the training set.

Severity-rich retinal & oculomic cohorts, including underrepresented and complex cases, linked where applicable to metabolic state and multi-signal correlations (Imaging + Labs + History)

Designed to reduce geographic and clinical underrepresentation across wide-spectrum Phenotypes, Socioeconomic Status, Disease Severity, and real-world care pathways.

Diversity & Phenotypic Coverage

Phenotypic Coverage

Severity-stratified labeling across disease stages and modalities, with cohort distribution transparency reported.

Population Representation

Geographic and population ratios quantified to support bias analysis, robustness testing, and external validity.

Socioeconomic Context

Relevant contextual variables included where available, normalized, and flagged for missingness.

Subgroup Coverage Metrics

Age, sex, and clinically relevant subgroup balance measured and transparently reported.

Regulatory-Grade
Retinal Data. At Scale.

Clinically adjudicated fundus and OCT
cohorts built for AI validation, bias analysis,
and regulated deployment.

Deploy Validation-Ready Datasets with Full Traceability

Each cohort includes severity stratification, provenance logs, and audit-ready documentation for high-stakes model development.

Support Diligence with Structured Documentation

EU AI Act–aligned dataset datasheets, inclusion criteria, QC statistics, and export manifests are included with every release.

Selected cohorts are available now, with custom build-to-spec programs and evaluation packs offered under NDA

Data Standards & Interoperability

01

Metadata Standards

Mapped to FHIR R4/R5, DICOM, SNOMED CT, and LOINC to maintain semantic consistency across imaging, laboratory, and clinical systems.

02

Annotation Quality

Multi-clinician annotations with documented inter-rater reliability (IRR) scoring and traceable validation logs.

03

Representation Metrics

Measured across phenotypic severity, population representation, and contextual socioeconomic variables where available.

04

Regulatory Documentation

EU AI Act–aligned datasheets and QMS documentation delivered with full provenance and audit-ready artifacts.

Request a
Technical Overview

We provide tailored cohort specifications and capability summaries for teams evaluating datasets
for model development, validation, or regulated deployment.