Text to image generators for anatomical illustrations: potential and limitations

Area of Science:

• Medical Illustration
• Artificial Intelligence in Medicine
• Anatomy Education

Background:

• Artificial intelligence (AI) has advanced rapidly, offering efficient methods for medical illustration and visualization.
• AI-powered tools can create custom illustrations and visualize complex medical and anatomical content cost-effectively.

Purpose of the Study:

• To evaluate the strengths and limitations of AI in generating anatomical illustrations.
• To assess the potential of AI tools for application in anatomy education.

Main Methods:

• Assessed ChatGPT, Microsoft Designer, Craiyon, and Image FX for generating illustrations of organs and genital systems.
• Evaluated images based on anatomical accuracy, educational utility, and visual quality using a five-point Likert scale.

Main Results:

• Image FX, ChatGPT, and Microsoft Designer scored higher (3.46, 3.40, 3.24) than Craiyon (1.86).
• Generated images often lacked anatomical accuracy despite good visual quality.
• Microsoft Designer faced policy restrictions on sensitive content, while others produced inaccurate images.

Conclusions:

• AI text-to-image generators have capabilities and limitations for anatomical illustration.
• Future AI models need integration with accurate anatomical databases and terminology for improved medical education applications.