A full-system pipeline for fundus imagery analysis.
HorusVue is built around a single conviction: the value of an ophthalmic AI system is not in any individual model, but in the integrity of the pipeline that surrounds it.
We handle the full path — biomedical marker detection, image preprocessing, semantic segmentation, and trained-model disease classification — as one coherent system, with auditability at every stage.
Clinicians work with images. HorusVue works with the questions clinicians are actually trying to answer.
Automated identification of biomedical markers across the retinal field — optic disc, macula, vessels, drusen, microaneurysms.
Illumination correction, vessel enhancement, and quality scoring — applied consistently so downstream models see consistent input.
Semantic segmentation of pathological and anatomical regions, with confidence maps and uncertainty quantification.
Trained models for diabetic retinopathy, AMD, glaucoma, and a configurable set of additional conditions.
Every diagnosis is accompanied by interpretable evidence — saliency, segmentation overlays, marker traces — so a clinician can verify the reasoning.
DICOM-aware, PACS-compatible, and engineered to slot into existing clinical workflows rather than replace them.
We work directly with research groups and clinical teams to integrate HorusVue into specific protocols and workflows. Briefings are free and unhurried.