Open Access Open Access  Restricted Access Subscription Access
Open Access Open Access Open Access  Restricted Access Restricted Access Subscription Access

Skin Lesion Classification Using Convolution Neural Networks


Affiliations
1 Department of ECE, ANUCET, ANU, India
2 Department of ECE, RVR & JC Engineering College, Guntur, A.P, India
     

   Subscribe/Renew Journal


Skin cancer is one of the deadliest disease found in humans. These skin cancers are of various types like Basal Cell Carcinoma(BCC), Melanoma, Nevus, Seborrheic Keratosis (SK), Squamous Cell Carcinoma (SCC). Some of the skin cancers can be identified visually, but in order to diagnose a skin cancer patient should have to undergo for a biopsy test and it takes a long time to diagnose. To overcome this an automated skin lesion classification system has to be developed. In this work, a basic architecture of the Convolution Neural Network(CNN) model is used to classify different skin lesions. The proposed model achieved better accuracy for SCC Vs SK, BCC Vs SK, Melanoma Vs Nevus and Melanoma Vs SK are 0.9741, 0.9867, 0.9506 and 0.9734 respectively for 25 epochs when compared to the other related works .

Keywords

Skin Cancer, Basal Cell Carcinomsa, Melanoma, Nevus, Seborrheic Keratosis, Squamous Cell Carcinoma, CNN, Training Accuracy, Validation Accuracy.
Subscription Login to verify subscription
User
Notifications
Font Size


Abstract Views: 303

PDF Views: 0




  • Skin Lesion Classification Using Convolution Neural Networks

Abstract Views: 303  |  PDF Views: 0

Authors

K. S. Rajasekhar
Department of ECE, ANUCET, ANU, India
T. Ranga Babu
Department of ECE, RVR & JC Engineering College, Guntur, A.P, India

Abstract


Skin cancer is one of the deadliest disease found in humans. These skin cancers are of various types like Basal Cell Carcinoma(BCC), Melanoma, Nevus, Seborrheic Keratosis (SK), Squamous Cell Carcinoma (SCC). Some of the skin cancers can be identified visually, but in order to diagnose a skin cancer patient should have to undergo for a biopsy test and it takes a long time to diagnose. To overcome this an automated skin lesion classification system has to be developed. In this work, a basic architecture of the Convolution Neural Network(CNN) model is used to classify different skin lesions. The proposed model achieved better accuracy for SCC Vs SK, BCC Vs SK, Melanoma Vs Nevus and Melanoma Vs SK are 0.9741, 0.9867, 0.9506 and 0.9734 respectively for 25 epochs when compared to the other related works .

Keywords


Skin Cancer, Basal Cell Carcinomsa, Melanoma, Nevus, Seborrheic Keratosis, Squamous Cell Carcinoma, CNN, Training Accuracy, Validation Accuracy.



DOI: https://doi.org/10.37506/v10%2Fi12%2F2019%2Fijphrd%2F192205