Related Experiment Video
Updated: Mar 19, 2026
06:34
SCAnED - An Open-source Skin Segmentation Macro for Semi-automated Cell and Nuclei Detection in Epidermal and Dermal Skin Compartments
Published on: August 8, 2025
649
Efficient segmentation of large-area skin images: a statistical evaluation.
D P Filiberti1, J A Gaines1, P Bellutta1
1Oregon Health Sciences University, Biomedical Information Communication Center, Portland, OR, USA.
Summary
This study developed an automated system for monitoring pigmented lesions, finding it as reliable and consistent as human experts. The system accurately identifies lesions, showing high correlation with human assessments.
Area of Science:
- Medical image analysis
- Computational dermatology
- Automated health monitoring systems
Background:
- The SPOTS project aims to automate the monitoring of pigmented lesions over time.
- This research focuses on developing and evaluating an automated system for lesion analysis.
Purpose of the Study:
- To introduce and tune parameters for an automated pigmented lesion monitoring system.
- To statistically evaluate the system's performance against human operators.
Main Methods:
- System parameters for image segmentation, classification, and region-growing were optimized using synthetic and real image data.
- Performance was evaluated by comparing automated lesion identification with manual tracings by multiple operators.
- Cross-validation and analysis of independent, matched, and doubly-matched data were employed.
Main Results:
- The automated system identified more objects than human operators but reported comparable average areas.
- High correlation was observed between the system's lesion identification and human operator assessments.
- The system demonstrated consistency comparable to human inter-operator reliability.
Conclusions:
- The developed automated system for pigmented lesion monitoring is as reliable and consistent as human experts.
- The system shows a tendency to report slightly smaller lesion areas compared to human operators.
- The findings support the use of this automated system for clinical applications.
