# AI diagnosis from medical images: Materials

Source page: https://www.materialinsights.org/technologies/ai_medical_imaging/tech-materials/

#### Materials it needs

The commodities and chemicals it needs per unit, its share of each one's demand and how critical each is
 Needs

- [Si Silicon](https://www.materialinsights.org/commodities/silicon/) criticality low The software runs on silicon chips in hospital computers, scanners and data centres.
- [Cu Copper](https://www.materialinsights.org/commodities/copper/) criticality low Computers and data centres that train and run the networks are wired with copper.

as of 2026-11-11. Software has no material content of its own; the links name the hardware it runs on. No source gives amounts per device, so criticality is the writer's judgement.

#### Critical materials

The materials whose supply could hold it back

- No material holds the software back; its hardware is ordinary computing equipment.
- The scarce input is data: thousands of images graded by specialists, such as the 128,000 retinal images graded by a panel of 54 ophthalmologists ^[2].

as of 2026-11-11

#### Sources

1. [Deep Learning for Detection of Diabetic Eye Disease (29 November 2016)](https://research.google/blog/deep-learning-for-detection-of-diabetic-eye-disease/) Google Research blog (L. Peng and V. Gulshan), 2016

## Other sections

- [Summary](https://www.materialinsights.org/technologies/ai_medical_imaging/index.md)
- [Overview](https://www.materialinsights.org/technologies/ai_medical_imaging/tech-overview/index.md)
- [Science](https://www.materialinsights.org/technologies/ai_medical_imaging/tech-science/index.md)
- [Evolution](https://www.materialinsights.org/technologies/ai_medical_imaging/tech-evolution/index.md)
- [Energy footprint](https://www.materialinsights.org/technologies/ai_medical_imaging/tech-energy-footprint/index.md)
- [Industries](https://www.materialinsights.org/technologies/ai_medical_imaging/tech-industries/index.md)
