AI-Ready Healthcare Data - MIDAS for India 

India's disease burden is vast. Most of it is detectable early. ARTPARK builds the data infrastructure and AI tools that make early detection possible, at the last mile, with the equipment that's actually there.

Foundation - MIDAS

MIDAS - Medical Information & Imaging DataSets for India

Building the foundational dataset layer that India's health AI needs to actually work.

AI models trained on international datasets don't perform accurately on Indian patients. MIDAS  addresses this at the foundation by collecting, curating, and disseminating standardized medical imaging datasets from institutions across India ensuring not just domestic representation but demographic representation.

Our multimodal datasets cover

  • Oral Cancer 

  • Breast Cancer

  • Diabetic Retinopathy

  • Meningioma

  • Glaucoma

Stats that count

  • 1 TB+ Clinical Data Available Now

  • 5 Disease Areas Covered

  • 10+ Partner Institutions Nationwide

Highlights

  • Showcased at AI Action Summit Paris, 2025 and India AI Impact Summit 2026  

  • Model: Hubs-and-spokes - multi-institution collection & standardisation

  • Access: Open to researchers and innovators via secure platform

Our Partners

Solving real problems with AI in healthcare

India's disease burden is growing. Its healthcare infrastructure needs tools that scale. Early detection saves lives. Yet for millions of Indians, it remains out of reach. 

We build AI for Society, supporting initiatives that close this gap, and serve India’s population and its underserved communities where it matters most. In line with this, we are equipping frontline health workers and improving early detection.

MIDAS for Oral Cancer Screening

India carries nearly one-third of the global oral cancer burden. 70% of cases are diagnosed at Stage III or IV, when survival rates fall below 37%. The gap is not treatment, it is early detection, especially at the last mile where specialists do not reach.

What drives our work?

MIDAS for Tuberculosis Screening

TB presentation varies by population, co-morbidity profile (including diabetes, which is endemic in India), and the prevalence of non-TB pulmonary disease. Models trained on Western data underperform on Indian X-ray data from district hospitals.

Working on AI for research and screening?

ARTPARK co-innovates with health systems, research institutions, and technology partners on screening tools built for India. Let's build together.