Research Article

Skin Disease Diagnosis Using Convolutional Neural Network

Abstract

Visual similarities have made diagnosing skin illnesses more difficult for medical professionals. Although melanoma is one of the most commonly- known kind of skin illness, some fatalities in the past few years have been attributed to other illnesses. A major obstacle to developing a robust automatic classification system is the lack of huge datasets. This paper presents a deep learning method for skin cancer diagnosis. CNN were trained using transfer learning to produce classifiers that are both hierarchical and simple that are capable of distinguishing between seven distinct types of moles. In order to enhance performance, data augmentation techniques were applied to HAM10000 dataset, This has an extensive set of dermatoscopic images used in this research. Findings show that the DenseNet201 network performs well on this task, The system incorporates a CNN model to forecast the skin illness among seven different moles. In addition to that it will display the result along with information about the disease and nearby hospital recommendation.