Author(s) :

M.uday kumar.

Volume/Issue :

Abstract :

Skin cancer has become the most commonly diagnosed cancer worldwide since the 1970s, with both melanoma and non-melanoma cases increasing steadily, particularly in Western countries. According to the World Health Organization, melanoma accounts for one-third of all cancer diagnoses. In the United States, one in five individuals is expected to develop skin cancer during their lifetime. Early diagnosis significantly improves the survival rate, yet differentiating between malignant and benign lesions remains a major clinical challenge. Conventional diagnostic methods often fall short due to the visual similarity of lesions and limited access to expert dermatologists. This study investigates the use of deep learning techniques, particularly Dense Convolutional Neural Networks (DenseNet), to classify skin lesions accurately. Traditional machine learning models such as K-Nearest Neighbors, Support Vector Machines, and Decision Trees yielded suboptimal results in terms of accuracy. By contrast, our DenseNet model achieved an accuracy exceeding 86.6%, highlighting its potential for automated and precise skin cancer detection. This approach can play a vital role in aiding early diagnosis and improving patient outcomes.

No. of Downloads :

0