The human body is covered with skin which acts as a protective organ between the inside and outside of the body. Hence, it is inevitable to attract more diseases than any other organ of the body. Most of the skin diseases timidly appear on the skin, which leads to great negligence as some normal scratch or something else by the patients. This utter negligence of the patients leads to a delay in visiting a dermatologist on time, which can result in pain or bad skin health for the longest period in the patient’s life. Hence, as a boon to this, deep learning techniques assist in the early detection of skin diseases based on the skin images of the patients. A considerably large number of researches are carried out to detect the skin diseases using neural networks to yield a good accuracy of more than 90% in most of the skin diseases. But still many skin diseases have poor accuracy around 80%, Hence, this paper majorly concentrates on unleashing the techniques of the existing work to identify the skin diseases and the detection techniques that are yielding poor detection accuracy.