IronSpot AI Monitor

Fortify Health

IronSpot AI Monitor

AI-powered image analysis for accurate and rapid iron content verification in fortified flour.

New Delhi, Delhi, India

IND

Nonprofit

WHO reports anaemia caused 50 million years of healthy life lost due to disability in 2019 alone. Major causes include dietary iron deficiency, thalassemia and sickle cell trait, and malaria, with Africa and parts of Asia bearing 71% of the global mortality burden and 65% of the disability-adjusted life years lost. In India, iron deficiency affects 69.5% of children and 0.5 billion people, leading to severe health and economic consequences. Fortifying wheat flour with iron offers a population-wide solution to combat anemia without requiring dietary changes. However, ensuring the correct iron levels in fortified flour is crucial for effectiveness, regulatory compliance, and consumer safety. Manual iron spot tests, while affordable ($0.10), are unreliable and prone to human error, making accurate interpretation difficult. In contrast, laboratory tests provide precise results but are costly ($60 per test) and take over two weeks. This delay forces millers to rely on expensive, slow tests to adjust fortification operations, hindering quality control. A reliable, cost-effective, and rapid testing method is essential to maintain proper iron levels in wheat flour and effectively combat anemia.

The proposed intervention aims to develop an open-source, turnkey solution incorporating a machine learning model to analyze chemical reaction images from iron spot tests. Using advanced image processing and regression techniques, the tool will accurately estimate the added iron content in samples, enabling rapid, accurate, and cost-effective quality control for flour fortification. The tool will consist of three key components: Backend application: Manages image processing, model training, and result classification. It also orchestrates downstream actions, such as reuploading low-quality images, sharing results, and guiding users on next steps. Front-end Software as a Service (SaaS): Provides an interface for SaaS and admin users, offering features such as an image-testing playground, data curation and labeling, and usage monitoring dashboards. Chat-flow for end users: Integrates with chatbots like WhatsApp to facilitate guided workflows, result sharing, and interactive troubleshooting. Both the machine learning model and application will be open-source and built modularly for seamless deployment on any cloud platform, with initial testing focused on Google Cloud Platform (GCP). The technology stack includes React, FastAPI, PostgreSQL, Caddy, and Typebot. This solution aims to enhance the efficiency and scalability of iron fortification quality control.

Currently Fortify Health partners with more than 135 mills across 14 states in India and the number of mill partnerships are scaling up at rapid pace. These mills currently produce ~35,000 metric tonnes of flour per month, estimated to reach more than 7 million beneficiaries monthly. The millers fortify the wheat with iron, and if they want to measure the amount of iron in the flour, they have to send the sample to a lab and wait for at least a week to get the results. With the tool, they will instead upload a picture to a platform that hosts the ML model, and they will receive the test results in minutes. They will use the test results to ensure that the wheat they are supplying has the regulated amount of iron fortification. Improved quality assurance with increased accuracy and speed in monitoring iron content in fortified flour will lead to a better compliance to fortification standards. This finally results in safer and more effective reduction of iron-deficiency anemia for target populations.