학술논문
An Efficient Diagnosis of Melanoma Skin Disease Using DenseNet-121
Document Type
Conference
Author
Source
2023 3rd International Conference on Technological Advancements in Computational Sciences (ICTACS) Technological Advancements in Computational Sciences (ICTACS), 2023 3rd International Conference on. :908-912 Nov, 2023
Subject
Language
Abstract
This research describes a Melanoma Skin Disease Detection system that use Convolutional Neural Networks (CNNs) to categorize skin photos into several illness categories. People of all ages are affected by skin problems, which are a major public health concern and can have disastrous consequences if neglected. The proposed method has three major steps: picture pre-processing, feature extraction using a pre-trained CNN model, and classification with a fully connected neural network. The pre-processing step includes image resizing, normalization, and augmentation. The process of feature extraction involves utilizing a preexisting CNN model, specifically Densenet-121, for obtaining distinctive attributes from the provided images of the skin. The classification stage employs a fully connected neural network to categorize the skin photos into several illness groups. Using the ISIC dataset, which includes images of skin suffering from various illnesses, the system is trained and tested. With an accuracy rate of 95%, the proposed method for identifying skin disorders surpasses currently available state-of-the-art techniques. The suggested method may improve the speed and accuracy of skin disease identification, giving dermatologists and other healthcare practitioners a practical and dependable answer.