Advancement of Oral Cancer Detection with EfficientResnet: A Hybrid Deep Learning Model

Oral cancer is a serious and rising hazard. It is the 6th most frequent cancer in the world, and the 3rd most prevalent in India. Squamous cell carcinoma is responsible for 90% of all oral malignancies, with a 5 -year survival rate of around 60%. However, if discovered early, the survival rate might exceed 90%. Unfortunately, the death rate has not improved in the last three decades, making it the most expensive cancer to treat. The most prevalent kind is oral squamous cell carcinoma (OSCC), which commonly develops as a result of mouth potentially malignant illnesses. Current diagnostic approaches, such as surgical biopsy and histopathologic examination, have limited accuracy and efficiency. This study looks into applying deep learning to enhance OSCC diagnosis. An Ensemble deep learning model based on EfficientNetB3 and ResNet50 was created by using a dataset of histopathology pictures. The model reached 98.3% accuracy, indicating its potential as a reliable diagnostic tool. This advancement marks a huge step ahead in the speedy and exact identification of OSCC, which could enhance patient out comes via proactive treatment and therapy.