A Review of Automatic Hair Removal in Dermoscopy Images: From Image Processing to Deep Learning

Area of Science:

• Dermatology
• Medical Imaging
• Computer Vision

Background:

• Dermoscopy enables early melanoma detection but hair artifacts hinder automated analysis.
• Hair removal is essential for accurate diagnosis but often overlooked as a preprocessing step.

Purpose of the Study:

• To provide the first comprehensive review of automatic hair removal techniques in dermoscopy images.
• To categorize and evaluate existing and emerging hair removal methods.

Main Methods:

• Literature review covering publications from 1990 to 2025.
• Categorization of techniques including conventional image processing and deep learning (DL) architectures.
• Evaluation of generative models and hybrid approaches for hair removal.

Main Results:

• Identified advancements in hair removal over three decades.
• Assessed the current state of research in automatic hair removal for dermoscopy.
• Highlighted the transition from traditional methods to DL-based solutions.

Conclusions:

• Automatic hair removal is a critical but underexplored area in dermoscopy image analysis.
• Deep learning approaches show significant promise for improved hair artifact removal.
• Further research is needed to address remaining challenges and advance clinical application.