Detection of selected skin diseases using AI
|
Full Text |
Pdf
|
|
Author |
Afia Akyaa Dabanka, Derrick Commodore and Percy Okae
|
|
e-ISSN |
1819-6608 |
|
On Pages
|
345-361
|
|
Volume No. |
20
|
|
Issue No. |
6
|
|
Issue Date |
May 15, 2025
|
|
DOI |
https://doi.org/10.59018/032547
|
|
Keywords |
melanoma, psoriasis, eczema, AI, convolutional neural network (CNN), skin diseases, VGG19.
|
Abstract
In this work, three skin diseases, namely melanoma, psoriasis, and eczema, are detected and diagnosed using the CNN machine learning technique. They are among the most prevalent health concerns worldwide, and in deprived areas, early detection is often hindered by limited access to specialized care. These conditions can lead to severe health issues, making a timely and accurate diagnosis crucial. Traditional diagnostic methods are time-consuming and subjective, highlighting the need for automated solutions. This work developed an AI-based system by doing a comparative analysis of the results of VGG16, VGG19, and InceptionV3 architectures of Convolutional Neural Network (CNN). The system was trained on a curated dataset from sources like DermNet and ISIC, with the VGG19 architecture producing the best result of 99 % accuracy in detecting melanoma, psoriasis, and eczema. It was closely followed by the VGG16 architecture, which achieved an accuracy of 95 %, whilst the InceptionV3 achieved an accuracy of 87 %. To give patients a closer view of automated tests and results, a user-friendly web application was developed. This enabled users to upload images and receive an automatic diagnosis, with a feature that recommends consulting a doctor if inconsistencies are detected. The web application aims to address challenges by providing a machine learning-based diagnostic tool that can be accessed remotely by users. Moreover, this work represents a significant advancement in the use of artificial intelligence for healthcare applications. The integration of machine learning and image processing technologies not only enhances the accuracy of diagnoses but also exemplifies the growing trend of AI-driven solutions in medical diagnostics.
Back