Multiclass Skin cancer Detection using Deep Learning Approach


  • S.R. Nalamwar, S. Neduncheliyan


Transfer Learning,Skin Cancer,ResNet50,CNN


Skin cancer is a disease which can spread very rapidly and very difficult to detect in the early stage due to the high similarity in the skin lesions symptoms. Dermatologist is needed more experience and accurate result analysis in the early stage so that the Dermatologist treat the people in the early stage to improve the life span of the patient. Our system is use to classify the image in the multiclass with high accuracy so that it can help the Doctors to treat their patients with accuracy results. The dataset used in this work include the images for the Various types of skin cancers like dermatofibroma, vascular lesion melanoma, melanocytic nevus, basal cell carcinoma, actinic keratosis, benign keratosis, and Squamous cell carcinoma .Melanoma is the very dangerous cancer which can spread very rapidly if it is not detected in the early stage. The proposed system we work on the ISIC 2019 Dataset and classify the images in the eight multiclass skin cancers. In this work we are going to use the Transfer learning ResNet 50 to train the model by considering the initial values of the parameters and changing its values through the transfer training. If the images those are not belongs to this eight classes are classify as unknown. Our system achieve the good  accuracy, Precision. Results shows our proposed system using CNN Transfer Learning it gives more accurate results as compare to the previous system.



How to Cite

S.R. Nalamwar, S. Neduncheliyan. (2022). Multiclass Skin cancer Detection using Deep Learning Approach. Computer Integrated Manufacturing Systems, 28(10), 400–408. Retrieved from