Simplifying The Complex Diagnosis of Nail Fungus Disease Detection Using An Deep Learning Hybrid Model

Nail fungus is a extensive contagious disease impacting millions world-wide, usually causing changes such as discolouration, nail thickening, and uneasiness. Nail diseases can indicate health issues, and initial detection is important for effectual treatment. Conventional diagnosis often depends on time-consuming laboratory tests or subjective clinical assessments. The use of technology to improve medical treatment, especially in third-world countries where medical resources are limited. The aim of our research is to identify effective vision-based medical image analysis methods and assessing hybrid models to achieve precise detection of nail fungus. The hybrid models combine machine learning and convolutional based approaches. The train and test data were classified into two categories: nail infected fungus disease and healthy well nails. Previous research are focused on individual single architectures such as CNN or MobileNet, our research developed a hybrid ensemble model that integrates seven single models, aiming to achieve remarkable performance on a huge nail fungus dataset. The final ensemble Hybrid model (include LinearSVM, KNN, Light CNN, MobileNetV2, Random Forest, Decision Tree, and EfficientNetB0), demonstrated best performance with an maximum accuracy of 99.70%, rather than all other single models. This research highlights an effectual nail fungus identification solution that extend diagnostic accuracy and minimize errors, improving overall healthcare facility through advanced deep learning based computer vision solution for effectual nail disease identification.