Existing equipment.
Start with compatible image exports from retinal cameras. Each camera, export format, and computer combination needs its own qualification.
← All productsRetinal AI,
Series 1 · Primary focus
EyeOps
Retinal AI,
closer to care.
Clinician-assist software for care settings with limited resources. We’re developing support around retinal cameras and computers already in use, starting with compatible image exports and a reviewable workflow.
Research prototype- 01Import an image
Begin with a supported retinal image export.
- 02Examine the evidence
View model results alongside the source image and available context.
- 03Keep people in control
A clinician reviews the evidence and decides the next step.
Illustrative workflow · Development direction
Designed for constrained settings
Work with what
care teams have.
A useful first step begins with the actual equipment and people in a care setting.
Local deployment in view.
We’re researching inference near the point of care, where compute and connectivity may be limited. Offline capability and low-power performance still need to be proven on target devices.
Reviewable assistance.
Keep images, supplied context, and model outputs together. The aim is to support a clinician’s review and preserve the evidence behind it.
Our path to a first pilot
Start with a workable setting.
Learn. Then expand.
We’re assessing resource-constrained care settings to find a practical first fit: accessible equipment, an engaged clinical team, and a workflow that can be evaluated clearly.
No pilot partnership or supported device list is confirmed yet.
- 01Understand the setting
Document the camera, export workflow, available computer, and connectivity.
- 02Qualify a narrow workflow
Evaluate image quality, runtime, model performance, and the clinician’s review process.
- 03Expand from evidence
Use the findings to inform support for additional devices and care settings.
A few useful distinctions
What to know.
Which cameras are supported?
A supported camera list has not been established. Current work starts with bounded JPEG and PNG image imports; an import working successfully does not prove that a camera or clinical workflow is supported.
Will EyeOps work entirely offline?
Local and edge deployment are part of the direction. We have not established fully offline operation on target equipment, so connectivity requirements remain part of device qualification.
Does it cover every retinal scan or health condition?
The current model scope is conventional color-fundus imaging and diabetic-retinopathy research. OCT, fundus autofluorescence, ultra-widefield imaging, and other conditions need separate models or evaluation.
Can it be used for clinical diagnosis now?
Clinical readiness has not been established. The prototype supports research and review; it does not provide autonomous diagnosis or replace professional clinical judgment.