ABSTRACT
Oral cancer is a severe public health hazard in the modern world, affecting mostly those over the age of 40 and increasing more widely with time. Oral squamous cell carcinoma (OSCC) is a kind of cancer that occurs in the mouth. It is the most prevalent cancer in the mouth, and it is the first and second most common illness in the two people who are afflicted by it, respectively. Dental cancer is a dangerous medical disorder that affects individuals all over the world, with an estimated 177,384 fatalities occurring each year in the United States alone. Oral sickness has been affecting two-thirds of the population in Asia for the past three decades, and the ailment is particularly prevalent in low- and middle-income nations, according to the World Health Organization. It is likely that increased use of robotics to distinguish between harmful and dangerous damage inside the oral cavity might result in lower costs and an earlier diagnosis of the illness. An oral lesion library with a bigger than normal selection of well-annotated oral lesions is essential. This article presents a neural network-based oral cancer detection system that makes use of a range of image processing methodologies. For this, we are using convolutional neural networks for the automated identification of oral lesions at the initial stage of oral cancer detection, which is the first step in the detection process. Following training, the system will be able to recognise the particular qualities of the items and compare them within a certain category depending on the information it has gathered about them. It is possible to construct a new output every time the specified inputs match up with the training data. Performance metrics, as well as the outcomes of image categorisation and object identification, are included in the output result, which is explored in more detail below.
