An Efficient Skin Cancer Diagnostic System Using Bendlet Transform and Support Vector Machine

Area of Science:

• Dermatology
• Medical Imaging
• Computer Science

Background:

• Melanoma, a dangerous skin cancer, has a high survival rate with early diagnosis.
• Accurate classification of dermoscopic images is crucial for timely melanoma detection.

Purpose of the Study:

• To develop a novel skin cancer classification (SCC) system using dermoscopic images.
• To evaluate the efficacy of Bendlet Transform (BT) as a feature extraction method for SCC.

Main Methods:

• Preprocessing involved median filtering to remove noise and hair from dermoscopic images.
• Feature extraction utilized Bendlet Transform (BT) for directional representation, classifying curvature, location, and orientation.
• Two Support Vector Machine (SVM) classifiers were implemented for the classification task.

Main Results:

• The SCC system employing Bendlet Transform demonstrated superior performance compared to other image representation systems.
• Bendlet Transform effectively captured relevant features for distinguishing skin cancer from benign lesions.

Conclusions:

• The developed SCC system using Bendlet Transform and SVM is a promising tool for accurate and early melanoma detection.
• Bendlet Transform offers a significant advantage over Wavelets, Curvelets, Contourlets, and Shearlets for skin cancer classification.