An Empirical Analysis of S-CNN based Oral Tongue Squamous Cell Carcinoma

Oral cancer refers to malignant tumors that begin in the mouth and spread to other parts of the oral cavity. It is a well-known and widely-recognized kind of human cancer. Risk in oral cancer can be reduced by not using paan, reducing alcohol consumption, not smoking, covering lower lip when you go outside, and getting the HPV vaccine. If oral cancer is caught in its early stages, OTC treatments can be effective. Automatic oral cancer detection paves the way for quicker diagnosis. Using existing software in a novel way, the authors of “A Novel Method for Automatic Detection of Early-Stage Oral Cancer” set out to create a system that is accessible to non-specialists but nonetheless very accurate in its early detection of mouth cancer. To accomplish these goals and boost image quality while decreasing noise, an anisotropic diffusion filter is utilized in pre-processing. Image segmentation is performed using Fuzzy C-Means (FCM). The GLCM (Grey Level Co-Occurrence Matrix) is used for this. Energy, homogeneity, contrast, and correlation are the extracted textural features. Preprocessing, image segmentation, feature extraction, and model training are all used in the proposed method. GF is utilized in the preprocessing stage. It employs PFLSCM for application in image segmentation. It then employs sparse CNN for model training after employing GLCM for feature extraction. A comparison is made between the suggested method and two others, including CNN and SVM. The proposed method trumps the competition.