IMPLEMENTASI TRANSFER LEARNING MENGGUNAKAN MODEL EFFICIENTNET B3 PADA KLASIFIKASI PIGMEN KANKER KULIT

Skin cancer has several causative factors such as genetic factors and excessive ultraviolet radiation from sun exposure. Despite the possibility of early diagnosis, people often hesitate to consult with experts, worsening their condition. Biopsy is a common technique for diagnosing skin cancer, but it is often complex and costly. According to the World Health Organization (WHO), skin cancer is one of the leading causes of death worldwide, with 10 million deaths recorded in 2020. With advancing technology, skin cancer can be detected using deep learning through the application of Convolutional Neural Network (CNN) methods using the transfer learning model EfficientNet-B3 for pigment classification. The dataset used is from the International Skin Imaging Collaboration (ISIC) 2018, consisting of 10015 skin dermoscopy images. The data augmentation and preprocessing process involves resizing images to 75 x 100 pixels and applying hair removal preprocessing. This study focuses on the classification of two classes, Benign and Malignant. Experimental results show that using the EfficientNet-B3 model successfully increases accuracy to 86%, with precision, recall, and F1-Score values of 0.86, 0.92, and 0.89 respectively when hair removal preprocessing is applied