Skin Disease Diagnosis by using Different Machine Learning & Deep Learning Approach

When a person has a specific skin problem, he or she must go to the doctor in person, which involves time and money. It may also be difficult to detect the issue if the affected individual is out of reach of doctors and hospitals, and some people may be hesitant to discuss their skin problem with a doctor. In order to deliver ac-curate treatment on time while reducing the doctor's efforts, a cost-effective, efficient, and time-consuming model must be designed. The proposed model will also allow the afflicted individual to express their problem by providing input in the form of a photo of the affected region to the skin cancer predictive model, which will then provide a Meta report that will be more useful in determining the kind of skin disease. This is mostly an image processing approach. Based on certain feature extraction, our suggested system can detect different types of skin disorders using CNN, Inception V3 and an SVM class. In this case, Acne, Melanoma, Psoriasis, Vitiligo, and Rosacea are five skin illnesses that are detected via different algorithms. After that, the pre-trained networks were fine-tuned.