A master's thesis at the College of Computer Science and Information Technology at University of Basrah, titled "Classification of Diabetic Tongue Images Using Deep Learning Techniques," explored this topic.
The thesis, submitted by student Ghazwan Hani Hussein, aimed to develop a diagnostic system for diabetes based on the analysis of tongue images using deep learning techniques. This approach draws upon the principles of traditional Chinese medicine, which links tongue characteristics to the overall health of a person's internal organs.
The study included the evaluation of several convolutional neural network models and the application of cumulative learning techniques such as majority voting, flexible voting, and stacking to enhance the predictive performance of the models. The study utilized two different datasets to increase the reliability of generalizations and enhance the generalizability of the models across diverse datasets.
