Extraction of specific parameters for skin tumour classification

Area of Science:

• Dermatology
• Medical Imaging
• Computer-Aided Diagnosis

Background:

• Melanoma is an aggressive skin cancer with increasing incidence.
• Early detection of melanoma is crucial for reducing mortality rates.
• Computer-aided diagnosis can assist general practitioners in identifying suspicious skin lesions.

Purpose of the Study:

• To develop a methodological approach for classifying skin tumors in dermoscopy images.
• To extract specific attributes for computer-aided diagnosis of melanoma.
• To differentiate between melanoma and benign skin lesions.

Main Methods:

• Preprocessing steps include hair removal and automatic image segmentation.
• Segmentation refines lesion boundaries using image edges.
• Extraction of asymmetry, border, color, and diameter (ABCD) attributes.
• An artificial neural network classifies lesions based on extracted ABCD features.

Main Results:

• The developed method effectively extracts diagnostic attributes from dermoscopy images.
• The ABCD features provide sufficient information to differentiate melanoma from benign lesions.
• The approach facilitates computer-aided diagnosis for melanoma detection.

Conclusions:

• The proposed methodology offers a robust approach for melanoma classification.
• This technique can aid in the early detection of malignant skin tumors.
• The computer-aided diagnosis system can support clinical decision-making for general practitioners.