Recent technical development of artificial intelligence for diagnostic medical imaging

Area of Science:

• Artificial Intelligence
• Medical Imaging
• Computer Vision

Background:

• Deep learning (DL) has spurred significant advancements in artificial intelligence (AI).
• Rapid progress in DL necessitates timely reviews to identify relevant technologies for medical applications.
• The fast-evolving nature of AI may render published medical AI technologies outdated quickly.

Purpose of the Study:

• To review state-of-the-art computer vision papers and presentations from 2018.
• To identify DL technologies with future clinical potential in medical imaging.
• To select technologies from a radiologist's perspective.

Main Methods:

• Review of computer vision literature and presentations from 2018.
• Focus on deep learning technologies.
• Selection criteria based on potential clinical impact in medical imaging.

Main Results:

• Identified key DL technologies with future clinical potential.
• Highlighted areas include generative adversarial networks, knowledge distillation, and general image datasets for supervised learning.
• The review emphasizes the need for continuous assessment of AI technologies in medicine.

Conclusions:

• Deep learning technologies from 2018 show promise for transforming medical imaging diagnostics.
• Radiologists can leverage these advanced computer vision techniques for improved diagnostic capabilities.
• Ongoing review and adoption of cutting-edge AI are crucial for medical advancements.