90%
Model Accuracy
30 Sec
Result Via Smartphone
100X
Lower Cost Than Lab Testing
200+
Mills Deployed
Industry
Food, Beverages, and Nutraceuticals
Business Function
Quality Assurance & Compliance
Capability
Computer Vision · Field Diagnostics · Program Monitoring
Tech Stack
Deep Learning (CNN) · WhatsApp API · CSV/JSON Integration
Overview
Fortify Health works with approximately 200 flour mills across 14 Indian states to support wheat flour fortification under India's national fortification mandate and the Government of India's Anemia Mukt Bharat initiative.
The solution integrates seamlessly into existing testing workflow.
To monitor iron fortification at this scale, program teams relied on manual iron spot tests and periodic laboratory analysis. Laboratory testing is expensive and slow for frequent monitoring, while manual interpretation of spot tests varied between operators and produced no digital record for program-wide analysis.
Hornbill Ag and Fortify Health co-developed IronLens, an AI-powered computer vision system that standardizes iron spot test interpretation using any smartphone. By delivering objective classifications, estimated iron concentrations, and instant digital records, IronLens enables frequent, scalable quality monitoring without changing existing field workflows or mill equipment.
Challenge
No Real-Time Visibility
Compliance data was delayed and aggregated, allowing substandard flour to leave mills before corrective action could be taken.
Subjective Field Testing
Iron spot test interpretation depended on operator judgement, producing inconsistent results with no documented record or network-wide visibility.
Cost & Infrastructure Constraints
Laboratory testing cost approximately $20 per sample with turnaround times of up to two weeks, making frequent monitoring impractical. Any solution also needed to operate on basic smartphones, without requiring dedicated hardware or reliable internet connectivity.
Our Solution
A computer vision model that standardizes the results that the human eye cannot.
IronLens is built around a convolutional neural network trained on a laboratory-validated image dataset paired with verified iron concentration metadata from nationally accredited laboratories. The model learns to distinguish the subtle color gradations of iron spot tests that separate compliant from non-compliant samples, differences that are inherently subjective when interpreted by the eye and can vary across operators, lighting conditions, and testing environments.
A quality officer simply photographs the test card using any smartphone. In under 30 seconds, IronLens classifies the sample as Low, Within, or High, aligned with FSSAI fortification standards. Alongside the classification, the platform returns an estimated iron concentration and confidence score, allowing program teams to monitor compliance trends over time rather than only identifying individual failures.
Every result is automatically timestamped and logged through WhatsApp, requiring no dedicated application or hardware. Results can also be exported in CSV or JSON format for integration into program dashboards and reporting systems.
Impact Delivered
90% model accuracy
Achieved with 94% recall on under-fortified samples, enabling reliable identification of non-compliant flour during routine field testing.
Large-scale program deployment
More than 500 quality officers now conduct 60–90 tests per mill each month, enabling continuous monitoring at scale.
100× lower testing cost
Reduced the cost of iron fortification monitoring from approximately $20 to $0.20 per test, making frequent compliance checks economically viable.
10M+ consumers reached
Supported large-scale flour fortification reaching more than 10 million consumers.
“This is exactly the kind of breakthrough we needed, a low-cost, smart, and scalable solution that works in the field, not just in theory.”
— Tony Senanayake, CEO, Fortify Health