A literature review of artificial intelligence (AI) for medical image segmentation: from AI and explainable AI to trustworthy AI

Zixuan Teng ¹, Lan Li ², Ziqing Xin ¹

⁰School of Future Science and Engineering, Soochow University, Suzhou, China.

Quantitative Imaging in Medicine and Surgery +

|

December 19, 2024

View abstract on PubMed

Summary

Explainable AI (XAI) and trustworthy AI (TAI) are advancing medical image segmentation. TAI offers a more reliable and safer approach for clinical applications, improving patient outcomes.

Area Of Science

Medical image analysis
Artificial intelligence in healthcare
Computer-assisted diagnosis

Background

Traditional medical image segmentation is labor-intensive and subjective.
Conventional AI improves efficiency but lacks transparency and predictability.
Explainable AI (XAI) and Trustworthy AI (TAI) address AI limitations in clinical settings.

Purpose Of The Study

To review the evolution of AI in medical image segmentation.
To highlight the development and impact of XAI and TAI.
To examine how XAI and TAI enhance AI systems for clinical use.

Main Methods

Literature synthesis of traditional, AI-based, XAI, and TAI segmentation methods.
Analysis of XAI principles for transparency and interpretability.
Examination of TAI's role in improving AI reliability, safety, and accountability.

Main Results

XAI enhances AI transparency but faces challenges in safety and robustness.
TAI builds upon XAI, offering a more reliable framework for medical image segmentation.
TAI integrates XAI principles with enhanced safety and dependability for clinical settings.

Conclusions

TAI presents a promising future for medical image segmentation.
TAI offers improved reliability and safety over conventional AI.
TAI can lead to better clinical outcomes and advance medical image processing.

Keywords:

Medical image segmentation artificial intelligence (AI) explainable AI (XAI) trustworthy AI (TAI)