RetGuard Screening System
A retinal camera, an AI processing unit, and a clinical interface — everything you need for multi-disease retinal screening in one compact package.

RetGuard Camera
High-resolution retinal imaging designed for point-of-care screening. Easy to use, quick to set up.

RetGuard Box
Compact AI processing unit. Plug in, power on, start screening. No configuration, no cloud setup, no IT overhead.

The Complete System
Retinal camera, RetGuard Box, and clinical interface — everything you need for multi-disease retinal screening in one integrated package.
Key Differentiators
One System, Five Conditions
No need for separate screening tools, separate subscriptions, or separate workflows. RetGuard screens for diabetic retinopathy, glaucoma, hypertensive retinopathy, AMD, and DME — all from a single retinal image, in a single session.
Completely Self-Contained
No cloud infrastructure. No internet dependency. No recurring API costs. All AI processing happens locally on the RetGuard Box, keeping patient data fully private and giving you complete control over your screening program.
Clinical-Grade Reports
Structured PDF reports are generated automatically with clear screening results, calibrated confidence levels, visual evidence, and referral recommendations. Ready for your clinical records, patient communication, or specialist handoff.
Specifications
| Category | Parameter | Detail |
|---|---|---|
| Compute | Processor | NVIDIA Jetson Orin NX 16 GB |
| Compute | AI Performance | 100 TOPS (INT8) |
| Compute | GPU | 1024 CUDA cores, 32 Tensor cores (Ampere) |
| Compute | CPU | 8-core Arm Cortex-A78AE @ 2.0 GHz |
| Compute | Memory | 16 GB LPDDR5 (102.4 GB/s) |
| Compute | Power | 10W–25W |
| AI Model | Parameters | 54.1M |
| AI Model | Input Resolution | 480 × 480 px |
| AI Model | Feature Dimension | 1,280-dimensional |
| Imaging | Modalities | Fundus photography + OCT |
| Imaging | Compatibility | Standard fundus cameras & OCT devices |
| Screening | Conditions | DR, Glaucoma, HTR, AMD, DME |
| Screening | Screening Time | < 30 seconds (all 5 conditions) |
| Screening | Output | PDF clinical report with Grad-CAM evidence |
| Screening | Integration | HL7-compatible |
| Connectivity | Interfaces | 3× USB 3.2 Gen2, GbE, WiFi, HDMI 2.1 |
| Connectivity | Cloud Dependency | None — fully offline operation |