DETECTION OF NON-MELANOMA SKIN CANCER BY DEEP CONVOLUTIONAL NEURAL NETWORK AND STOCHASTIC GRADIENT DESCENT OPTIMIZATION ALGORITHM

Nowadays, people are doing a lot of work outside for a living. When they roam outside, there may be a chance to enter types of bacteria or fungi in our bodies through the skin by either the polluted gases from the vehicles or the ultraviolet rays emitted by the Sun. The expansion of skin problems f...

Full description

Saved in:
Bibliographic Details
Main Authors: Premananda Sahu, Srikanta Kumar Mohapatra, Prakash Kumar Sarang, Jayashree Mohanty, Pradeepta Kumar Sarangi
Format: Article
Language:English
Published: Institute of Mechanics of Continua and Mathematical Sciences 2025-01-01
Series:Journal of Mechanics of Continua and Mathematical Sciences
Subjects:
Online Access:https://jmcms.s3.amazonaws.com/wp-content/uploads/2025/01/20061339/jmcms-2501005-DETECTION-OF-NON-MELANOMA-SKIN-CANCER-PS-SK-1.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Nowadays, people are doing a lot of work outside for a living. When they roam outside, there may be a chance to enter types of bacteria or fungi in our bodies through the skin by either the polluted gases from the vehicles or the ultraviolet rays emitted by the Sun. The expansion of skin problems for human beings has emerged as a significant problem, and the successful investigation has been observed as an arduous task for clinical experts or dermatologists. This paper has furnished an automatic diagnosis of skin cancer earlier with the help of deep learning techniques and the skin-related images captured by the Skin Biopsy test. In this approach, we detected non-melanoma using ensemble techniques related to deep convolutional neural networks and the stochastic gradient descent optimization technique. Furthermore, we used HAM 10000 as the data set for training and testing purposes, as well as the feature extraction technique Principal Component Analysis. This work also investigated a comparison of previous models. It was found that the proposed model gained an approximation of 98.57 % classification accuracy.
ISSN:0973-8975
2454-7190