Lesion detection & segmentation
Automatically detect and segment hemorrhages, exudates, neovascularization, and retinal-layer abnormalities from OCT and fundus images.
We provide end-to-end custom development of ophthalmology AI using OCT and fundus cameras — from problem framing and PoC through model development, validation, and SaMD readiness. From lesion detection and disease classification to severity grading and finding visualization (XAI), we build to fit your challenge.
Ophthalmology AI custom development is a service in which a specialist company designs, builds, validates, and operates AI — for lesion detection, disease classification, severity grading, and finding visualization on ophthalmic images such as OCT and fundus photographs. Building on our own ophthalmology AI (OCT retinal-disease classification and diabetic-retinopathy diagnosis), Kewel develops it together with our in-house specialist physicians so it is genuinely useful in clinical practice.
Modalities
Target diseases / findings (examples)
Targets are adjusted to your challenge.
Automatically detect and segment hemorrhages, exudates, neovascularization, and retinal-layer abnormalities from OCT and fundus images.
Beyond presence/absence, we build models for severity and stage grading, such as diabetic-retinopathy staging (e.g., Davis classification).
Automatically extract OCT findings with color-coding and tags, and visualize the diagnostic rationale with explainable AI (XAI).
From data design and annotation to dataset construction, under the supervision of our in-house specialist physicians.
We can start from problem scoping and a PoC. Please feel free to get in touch.