Comparative evaluation of AI algorithms for dermoscopic image analysis in skin disease diagnosis: a systematic review of model accuracy and methodological heterogeneity

Key Points

• Diagnosis accuracy for skin diseases varies significantly among AI algorithms, indicating uneven performance.
• Key metrics show that some AI models achieve up to 90% accuracy, while others are less reliable.
• Systematic review of multiple AI algorithms for dermoscopic image analysis was conducted to evaluate performance.
• Findings highlight the need for standardization in AI models to improve diagnostic reliability across platforms.