Skin disease detection using artificial intelligence in healthcare industry

Global ailments like acne and melanoma have become increasingly common amongst people everywhere due to things like air pollution and aspects of climate change. Understanding and diagnosing them early is very essential in averting serious problems later in life. The goal of this research is to create an AI-based system that will facilitate early skincare diagnostics and tailor preventive strategies to individual needs. Our work includes building and analyzing image datasets for training ML and DL models for classifying skin diseases. Several algorithms like SVM, Random Forest, KNN, and Convolutional Neural Networks (CNN) are utilized to find the optimal model for precise identification. Moreover, various VGG, ResNet, DenseNet, GoogleNet, MobileNet, EfficientNet, Xception, and other deep learning models are tested to determine whether they outperform CNN-based architectures. Various optimizers such as SGD, Lion, Lamb, and several others have been tested to improve the results for the classification problem. The dataset used in this research contains over 17,500 samples of eight distinct skin conditions.