Classification of Oral Squamous Carcinoma Histopathological images using Alex Net

Oral cancer may be indicated by a growth or sore in the mouth that does not heal. Oral cancer affects 50,000 Americans annually, 70% of whom are men. Oral cancers include those of the lips, tongue, cheeks, floor of the mouth, hard and soft palates, sinuses, and pharynx (throat). If not identified and treated right away, it could be fatal. Early detection makes treating oral cancer considerably simpler for medical professionals. However, most people receive a diagnosis when their ailment has progressed too far to be adequately treated. One of the most commonly occurring malignant neoplasms worldwide is oral cancer. Various risk factors are related to the event of carcinoma. Since cancers are deadly, early diagnosis of cancer is very important for initiating the treatment. Manual methods like a biopsy is time-consuming as it takes 2 to 3 days for obtaining a biopsy result. Physical screening of cancer with CT and x-ray images by the doctors is also less accurate. It was shown how deep learning with convolutional neural networks (CNNs) delivers effective detection and has had considerable success in classification. Several neuron-wise and layer-wise techniques were used with a CNN that had been trained using a publically accessible dataset of oral histopathology imaging data. In order to diagnose diseases, it was shown that neural networks can capture the colors and textures of lesions unique to those diseases, which is similar to human decision-making. So, using convolutional neural networks, we provide a method for classifying oral squamous cell cancer from images of oral histology. Here, we employ a selective regularisation technique to solve the overfitting issue that occurs when deep networks are trained on small amounts of data. The system's effectiveness is contrasted with that of other cutting-edge techniques.