Deep Learning Based Skin Lesion Detection with Interactive Chatbot

Abstract

One of the most common types of cancer that is reported in the world is skin cancer and is a significant health issue. The rising cases of skin cancer have been linked to many factors, which include exposure to ultraviolet radiation, environmental as well as modern lifestyle. Melanoma is a dangerous form of skin cancer, and one of the health-related issues of primary concern. The reason behind this is that melanoma is a serious form of skin cancer that is highly contagious and hence the high rate of mortality. Thus, early diagnosis of melanoma stands a chance of being diagnosed, and survival may become possible. Nonetheless, the conventional techniques of melanoma diagnosis such as visual examination and dermoscopy could be conducted only by a dermatologist. Even though conventional ways of diagnosing melanoma are effective, they may be subjective, time consuming and may require a dermatologist. Yet, since artificial intelligence was invented, there are other solutions to diagnosing melanoma. As an example, deep learning algorithms, in particular, CNN, can work well in the diagnosis of melanoma. This is so because deep learning algorithms, particularly CNN, are capable of detecting significant attributes of images, such as color, texture, and shape without the need of any human intervention. This plays a key role in the diagnosis of melanoma because slight variations play a vital role.