Skin Lesion Classification: A CNN Way 
Prasad Thakar1, Siddhivinayak A Kulkarni2
1Prasad Thakar*, School of Computer Engineering and technology, MIT World Peace University, Pune, Maharashtra, India.
2Dr. Siddhivinayak A Kulkarni, School of Computer Engineering and technology, MIT World Peace University, Pune, Maharashtra, India.
Manuscript received on April 11, 2020. | Revised Manuscript received on May 15, 2020. | Manuscript published on June 30, 2020. | PP: 274-278 | Volume-9 Issue-5, June 2020. | Retrieval Number: E9514069520/2020©BEIESP | DOI: 10.35940/ijeat.E9514.069520
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Skin lesion growth of unwanted cells on the upper most layer of skin. These lesions may conation cancerous cells which may lead to health issues to the patient and in severe cases may lead to patient’s demise. Dermatologists identify type of skin cancer by identifying it in image generated using dermatoscope and procedure known as Dermatoscopy. Previously there have been many studies which show classification of these dermatoscopic images using machine learning and deep learning solutions. Machine learning approaches use image processing techniques for identifying mole in given image and then for classification researchers have used techniques like SVM , random forest etc. With advances in field of deep learning there have been various methods proposed on classification of using CNN which achieves more precision and accuracy. In this paper we are proposing a CNN based approach for image classification with best overall accuracy of 78.08% and good multiclass AUC for all classes in HAM10000 dataset.
Keywords: Artificial Intelligence, Computer Vision, Deep Learning, Image Classification, Skin lesion classification, Convolutional-Neural-Network.