Related Experiment Video
Updated: Oct 10, 2025
13:44
Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
Published on: August 30, 2013
43.1K
Quantitative Comparison of Color Asymmetry Features for Automatic Melanoma Detection.
Summary
Simple brightness features best predict lesion asymmetry and melanoma in images. This aids researchers in developing more accurate automatic melanoma detection systems.
Area of Science:
- Dermatology
- Computer Vision
- Medical Imaging Analysis
Background:
- Melanoma detection relies on asymmetry assessment.
- Color asymmetry features are utilized for automatic melanoma detection from images.
Purpose of the Study:
- Compare color asymmetry features for melanoma detection.
- Evaluate feature accuracy in predicting lesion asymmetry and differentiating melanoma from benign lesions.
Main Methods:
- Evaluated nine color asymmetry features.
- Used a dataset of 277 lesion images.
- Assessed accuracy in predicting asymmetry and melanoma differentiation.
Main Results:
- Simple features based on brightness difference between lesion halves performed best.
- These features accurately predicted asymmetry and melanoma.
Conclusions:
- Brightness-based asymmetry features are effective for melanoma detection.
- Findings assist researchers in selecting optimal features for improved automated systems.
- Potential to reduce clinician workload by pre-screening benign cases.
