Efficient Skin Cancer Detection Using Enhanced Adaptive Hybrid Rnn Classification For Bio Medical Application

Abstract

Skin cancer is the uncontrolled multiplication of skin cells due to irreparable DNA damage. Recognizing skin cancer in images is a difficult medical undertaking. For the best possibility of a successful treatment outcome, skin cancer must be detected early. Early skin cancer screening is now achievable because of technological advancements. Dermoscopy is used in computer assisted diagnostics to capture the skin image. The skin image is initially pre-processed in this paper. Using an image segmentation technique, the lesion part is first segmented following pre-processing, and then certain features are recovered from the segmented lesion. After features are extracted, the skin image is classified as either healthy skin or melanoma skin cancer using a deep learning method called Enhanced Adaptive Hybrid RNN (EAH-RNN). The experimental findings of the suggested system prove that EAH-RNN provides the highest accuracy.