Background/Objectives: Coated tongue is a common oral condition with notable clinical relevance, often overlooked due to its asymptomatic nature. Its presence may reflect poor oral hygiene and can serve as an early indicator of underlying systemic diseases. This study aimed to develop a robust diagnostic model utilizing convolutional neural networks and machine learning classifiers to improve the detection of coated tongue lesions. Methods: A total of 200 tongue images (100 coated and 100 healthy) were analyzed. Images were acquired using a DSLR camera (Nikon D5500 with Sigma Macro 105 mm lens, Nikon, Tokyo, Japan) under standardized daylight conditions. Following preprocessing, feature vectors were extracted using CNN architectures (VGG16, VGG19, ResNet, MobileNet, and NasNet) and classified using Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Multi-Layer Perceptron (MLP) classifiers. Performance metrics included sensitivity, specificity, accuracy, and F1 score. Results: The SVM + VGG19 hybrid model achieved the best performance among all tested configurations, with a sensitivity of 82.6%, specificity of 88.23%, accuracy of 85%, and an F1 score of 86.36%. Conclusions: The SVM + VGG19 model demonstrated high accuracy and reliability in diagnosing coated tongue lesions, highlighting its potential as an effective clinical decision support tool. Future research with larger datasets may further enhance model robustness and applicability in diverse populations.

Aim: This study aimed to perform clinical diagnosis and treatment planning of mucous retention cysts with high accuracy and low error using the deep learning-based EfficientNet method. For this purpose, a hybrid approach that distinguishes healthy individuals from individuals with mucous retention cysts using panoramic radiographic images was presented. Material and Methods: Radiographs of patients who applied to the Department of Oral and Maxillofacial Radiology, Faculty of Dentistry, Fırat University between 2020 and 2022 and had panoramic radiography for various reasons were evaluated retrospectively. A total of 161 radiographs, 82 panoramic radiographs with mucous retention cysts and 79 panoramic radiographs without mucous retention cysts, were included in the study. In the classification process, deep feature representations or feature maps of the images were created using eight different deep learning models of EfficientNet from B0 to B7. The efficient features obtained from these networks were given as input to the support vector machine classifier, and healthy individuals and patients with mucous retention cysts were classified. Results: As a result of the model training, it was determined that the EfficientNetB6 model performed the best. When all performance parameters of the model were evaluated together, the accuracy, precision, sensitivity, specificity, and F1 score values were obtained 0.878, 0.785, 0.916, 0.857, and 0.846, respectively. Conclusion: The proposed hybrid artificial intelligence model showed a successful classification performance in the diagnosis of mucous retention cysts. The study will shed light on other future studies that will serve the same purpose.

Objective: One of the serious mental disorders in which people interpret reality in an abnormal situation is schizophrenia. A combination of extremely disordered thoughts, delusions, and hallucinations occurs due to schizophrenia, and the person's daily functions are seriously impaired due to this disease. For general cognitive activity analysis, electroencephalography signals are widely used as a low-resolution diagnostic tool. This study aimed to diagnose schizophrenia using the transfer learning method by including the EEGs of 73 patients diagnosed with schizophrenia, and 67 patients from the healthy group. Material and Method: In the first step of the study, digital electroencephalography signal data was converted into spectrograms to make them usable. In the classification phase, ResNet18, ResNet50 and EfficientNet models, which are FastAI, and Convolutional Neural Network (CNN) based deep learning models, were used. Results: Despite the complexity of electroencephalography data, CNN-based models in the study were successful in capturing different aspects of neurophysiological activity. The best performance was obtained from the ResNet-50 model with an accuracy rate of 97%. Afterwards, the classification process was finalized with 95% ResNet-18, and 83% EfficientNet models, respectively. Conclusion: It is thought that the classification performance of the result obtained in the application is promising, and may be a guide for future studies.