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.
