Retinopathy screening with Farsi NLP

An on-premise workflow that analyses fundus images and Persian clinical notes to triage diabetic retinopathy cases in Semnan clinics.

Slide 1

Diabetic retinopathy screening using computer vision models and lesion heatmaps.

Slide 2

Clinical pipeline merging image QC, Persian summarisation and HIS/PACS integration.

Slide 3

Automated Persian SOAP note summarising key indicators and referral urgency.

Project gallery1/3

Digital health

Retinopathy screening with Farsi NLP

An on-premise workflow that analyses fundus images and Persian clinical notes to triage diabetic retinopathy cases in Semnan clinics.

StatusOperational

ProvinceSemnan

DomainDigital health

Executive summary

Diabetic populations in Semnan continue to grow. The screening platform fuses retinal imaging with Persian-language notes to flag high-risk patients and support clinicians.

Challenges addressed

Limited specialist availability and long waiting lists.
Unstructured Farsi notes that hide relevant risk factors.
Strict privacy requirements mandating on-prem deployment.

Delivery approach

Apply computer vision to detect lesions in fundus images.
Use NLP pipelines to extract risk indicators from clinical text.
Score cases and populate referral dashboards.

Delivered outcomes

Persian-language screening system with audit trails.
Clinical summaries for specialists and health authorities.
Data governance artefacts covering privacy and retention.

Book a consultation Back to projects