The College of Computer Science and Information Technology at University of Basrah organized a scientific lecture titled "Diabetic Tongue Image Classification Using Deep Learning Techniques"
The lecture aimed to classify non-invasive diabetes mellitus from tongue images. Drawing on traditional Chinese medicine principles that link tongue shape to systemic health, this study evaluates the effectiveness of convolutional neural networks (CNNs) in detecting indicators related to diabetes.
The lecture, delivered by graduate student Ghazwan Hani Hussein in the Department of Computer Science, involved the use of two distinct tongue image datasets to train and evaluate different convolutional neural network architectures. Additionally, ensemble learning methods, including majority voting, smooth voting, and cluster stacking, were applied to improve classification performance. This was done to improve model generalization and overfitting.
