Oral cancer is one of the most common and deadly forms of malignancy globally, with high morbidity and mortality rates, often due to late diagnosis. Early detection plays a crucial role in improving survival rates, making it essential to develop accurate and efficient diagnostic tools. This research proposes an AI-driven hybrid system for the diagnosis of oral malignancies using machine learning and deep learning algorithms. The system employs Convolutional Neural Networks (CNN), DenseNet, MobileNet architectures, and a Recurrent Neural Network (RNN) to classify images of oral lesions, aiming to predict the presence of oral cancer with high accuracy. The CNN is used for extracting hierarchical image features, DenseNet for efficient feature reuse, MobileNet for real-time application without compromising accuracy, and RNN for temporal pattern recognition and sequence learning in image-based data. The model is trained on a dataset of labeled oral lesion images, and its performance is evaluated using various metrics such as accuracy, sensitivity, specificity, and F1-score. The results demonstrate the potential of hybrid deep learning models in the early diagnosis of oral malignancies, paving the way for the development of AI-powered diagnostic tools that can be integrated into clinical settings for improved patient outcomes.

Keywords:

Oral Cancer, Machine Learning, Deep Learning, Convolutional Neural Networks (CNN), DenseNet, MobileNet, Recurrent Neural Network (RNN), Oral Malignancies, Early Detection, Image Classification, Medical Imaging, Cancer Diagnosis, AI-driven Diagnostics, Predictive Modeling, Feature Extraction, Medical AI, Hybrid Model, Temporal Analysis.