AI glossary · Basic · Computer vision · Healthcare & bio
Medical imaging AI
Medical imaging AI is technology in which AI analyzes medical images such as X-rays, CT, MRI and ultrasound to find lesions (areas suspected of disease) or assist in reading the images. The final diagnosis is usually made by a doctor, and the AI acts as a "second pair of eyes."
About this term
- Level
- Basic · Technical terms you often meet in AI news
- Category
- Industry AI
- Fields
- Computer vision
- Industries
- Healthcare & bio
- Added
- 2026-10-10
How it works
An AI model is trained on a large number of medical images that doctors have annotated in advance. Deep learning models such as CNNs have long been used for image processing, and transformer-based models are also used today. When the trained model receives a new image, it marks the location of suspicious areas, gives a score for the likelihood of an abnormality, or draws the boundaries of organs to measure their size. These results are displayed in the hospital's image storage and viewing system for doctors to consult when reading images.
How medical imaging AI assists image reading
- 1Image acquisitionX-ray, CT, MRI, etc.
- 2AI analysisMarks suspicious areas & scores
- 3Shown on reading screenDisplayed with the image
- 4Doctor's readingChecks with AI results as reference
- 5Final diagnosisDecided by the doctor
Why it matters
The number of images taken at hospitals keeps growing, but the number of specialists who can read them is limited. When AI picks out urgent images first or marks small lesions that are easy to miss, it helps improve reading speed and consistency. It can also serve as an aid for reducing variation in readings caused by fatigue or individual differences. For these reasons, medical imaging is considered one of the first fields where AI entered actual clinical practice.
Things to know
Performance can drop with equipment, imaging conditions or patient populations that differ from the hospital where the model was trained. That is why it is important to validate the model again on the hospital's own data before actual deployment. In many countries, AI used for diagnosis is classified as a medical device and goes through an approval process. Also, it is often hard to explain why the AI flagged a particular area, so doctors need to check the results rather than trust them as-is.
Example
Chest X-ray reading assistance services mark suspicious areas such as lung nodules or pneumothorax in color on the image and show a score for the likelihood of an abnormality. Products that assist in reading mammograms, or that pick out brain MRI images suspected of stroke first and alert medical staff, are also used in many hospitals. Most of them display results within the doctor's reading screen, and the doctor makes the final judgment.