Feasibility of Artificial Intelligence-Based Image Enhancement Program for Anatomical Dissection Photographs

Area of Science:

• Medical Education
• Anatomical Imaging
• Artificial Intelligence

Background:

• Anatomical photographs are crucial for medical education and research.
• Dissection images require high detail for accurate visualization.
• Assessing AI's role in enhancing anatomical image quality is needed.

Purpose of the Study:

• To investigate the feasibility of AI for enhancing anatomical dissection photographs.
• To determine if AI enhancement improves subtle visual details under magnification.
• To evaluate the impact on anatomical accuracy and training value.

Main Methods:

• 50 anatomical photographs (2001-2024) were enhanced using Upscayl (16× REAL-ESRGAN).
• Original and enhanced images were compared at 1× to 20× magnification by 40 experts.
• Qualitative assessment focused on anatomical accuracy, noise, edge definition, and training value.

Main Results:

• AI enhancement yielded subtle visual differences, more apparent in deep anatomical regions at higher magnifications.
• Improvements in edge definition and noise reduction were noted, particularly for lower-resolution images.
• No anatomical distortions were introduced, but file size increased significantly.

Conclusions:

• AI-based image enhancement is feasible for anatomical dissection photographs.
• Modest visual benefits are possible, especially for older/lower-resolution images at higher magnification.
• Further optimization is needed to reduce file size and processing time for routine use.