Combined empirical mode decomposition and texture features for skin lesion classification using quadratic support...

Area of Science:

• Dermatology and Medical Imaging
• Computer Science and Machine Learning

Background:

• Basal cell carcinoma (BCC) is a prevalent malignant skin lesion.
• Accurate diagnosis is crucial to minimize errors and improve patient outcomes.
• Automated image analysis offers a promising approach for objective lesion classification.

Purpose of the Study:

• To develop and evaluate a novel automated technique for classifying skin lesions.
• To differentiate between malignant Basal cell carcinoma and benign nevus using image processing.
• To enhance diagnostic accuracy for skin lesions through computational methods.

Main Methods:

• A hybrid feature extraction method combining bi-dimensional empirical mode decomposition and gray-level difference statistics was employed after hair removal from lesion images.
• The extracted hybrid features were utilized for classification using a quadratic support vector machine (Q-SVM).
• The system was evaluated on its ability to classify skin lesions into Basal cell carcinoma and nevus categories.

Main Results:

• The proposed system demonstrated exceptional performance, achieving 100% accuracy.
• The method also attained 100% sensitivity and 100% specificity in classifying skin lesions.
• Performance was superior compared to other support vector machine approaches and alternative feature extraction methods.

Conclusions:

• The developed Q-SVM model effectively classifies Basal cell carcinoma using the proposed hybrid feature set.
• The automated system shows high potential for accurate and reliable skin lesion diagnosis.
• This approach offers a valuable tool for dermatologists in identifying malignant skin lesions.