Classification of skin cancers color image using hybrid feature extraction with ANN classifier

What is it about?

This article focuses on the development of a hybrid system for the high-accuracy classification of skin lesions. It emphasizes the importance of early detection of skin cancer, both benign and malignant, to ensure timely medical intervention and safeguard patients' lives. The proposed system consists of three main stages: the extraction of the region of interest using pre-processing techniques, feature extraction using multiple methods (Contrast, DCT, DWT, Gradient), and the classification of skin lesions using an artificial neural network. The system is evaluated using the HAM10000 Dataset 2018, and its performance achieves high classification accuracy across various types of cancerous skin lesions, demonstrating the effectiveness of the hybrid approach.

Featured Image

Why is it important?

This work is important because it presents a hybrid system that combines multiple feature extraction techniques (Contrast, DCT, DWT, Gradient) for the high-accuracy classification of skin lesions, a crucial aspect in early skin cancer detection. By utilizing the HAM10000 Dataset 2018, the system achieves impressive classification accuracy rates of up to 99.8%. Early and accurate detection can significantly improve patient outcomes, making this research timely and impactful, particularly in advancing non-invasive diagnostic technologies. The proposed hybrid approach could serve as a model for future automated systems in medical image analysis, improving healthcare efficiency and accuracy worldwide.

Perspectives