Early Detection Of Skin Cancer Using Convolution Neural Network

Abstract

Skin is the most powerful protection of important 
organs in the human body. It acts as a shield to 
protect our internal body to get damaged. But this 
important part of the human body can be affected 
by so serious infections caused by some fungus or 
viruses or even dust too. Around the world, 
millions of people suffer from various skin diseases. 
From acne problems to eczema people suffer a lot. 
Sometimes a small boil on the skin can turn into a 
severe issue or even an infection that will cause a 
major health issue. Some skin issues are so 
contagious that one can be affected by another just 
with a handshake or using a handkerchief. A 
proper diagnosis can result in proper medication 
that can reduce the miseries of the people 
suffering create awareness. In this project we are 
using CNN (convolution neural networks) to 
classify skin diseases from images as CNN gain 
lots of success and popularity in the field of image 
classification. To train CNN we have used skin 
disease dataset which contains 9 different types of 
diseases such as 'Actinic Keratosis','Basal Cell 
Carcinoma', 'Dermatofibroma', 'Melanoma', 'Pigmented 
'Seborrheic 
Keratosis', 
Benign 
Keratosis', 
'Squamous 
Cell 
Carcinoma' and 'Vascular Lesion'. After training 
CNN algorithm we can upload any test image then 
CNN will detect and classify disease from that 
image.