Tongue disease recognition using deep learning

Abstract

The tongue is extensively involved in several vital functions, such as taste, speech, eating, swallowing, and oral hygiene. the importance of maintaining the health of the tongue, this "Tongue Disease Recognition Using Deep Learning" research project. We investigated the use of sophisticated convolutional neural networks (CNNs) in the identification and categorization of different tongue disorders using medical photographs. I assessed how well a number of deep learning models performed— InceptionV3, MobileNet, ResNet50, and VGG16—in diagnosing ailments including Median Rhomboid, Black Hairy Tongue, Tongue Ulcer, Ankyloglossia, and Geographic Tongue.These models were trained, verified, and tested with 10,598 highresolution tongue photos from a comprehensive dataset to guarantee reliable performance. The result of the study provided that the MobileNet model had the best accuracy, 93.83%, proving that it is better able to identify illnesses of the tongue. This study demonstrates how deep learning technology can revolutionize medical diagnostics by offering a non-invasive, effective, and precise way to identify diseases early on.In order to maintain regulatory compliance, ethical issues such as patient privacy and data security were carefully taken into account. This study emphasizes how important it is to include deep learning models into clinical practice since they have the potential to improve healthcare accessible, especially in rural and underdeveloped areas.In order to further increase the system's usability in clinical situations, future work will concentrate on growing the dataset, investigating more complex designs, and improving the system's real-time diagnostic capabilities. This study shows that applying AI to medical diagnostics is feasible and advantageous, with the potential to enhance healthcare outcomes and boost accessibility.