Classification Models for Skin Tumor Detection Using Texture Analysis in Medical Images

Area of Science:

• Dermatology
• Medical Imaging
• Computer Science

Background:

• Medical imaging is crucial for early disease diagnosis.
• Accurate identification of skin lesions like melanoma and nevus is essential for timely treatment.

Purpose of the Study:

• To develop and evaluate a novel strategy for analyzing skin images to model, classify, and identify skin lesions.
• To assist dermatologists in the early diagnosis of melanoma through automated image analysis.

Main Methods:

• Utilized machine learning algorithms applied to skin image data.
• Extracted features including first and second-order statistics, Gray Level Co-occurrence Matrix (GLCM), keypoints, and color channel information (RGB, grayscale).
• Compared various mathematical classifier models to determine the most effective for lesion identification.

Main Results:

• The study successfully characterized decisive information from skin images for classification.
• Identified key features and models that enable accurate differentiation between melanoma and nevus.
• Demonstrated the potential of the proposed strategy in aiding dermatological diagnosis.

Conclusions:

• The developed machine learning strategy offers a promising approach for the automated analysis and classification of skin lesions.
• This method can significantly support dermatologists in the early and accurate identification of melanoma.
• Further research can refine these techniques for improved diagnostic accuracy and patient outcomes.