A Cloud Based Approach for the Classification of Oral Cancer Using CNN and RCNN
Tanala S Ch U S R Bhupathia, S V V D Jagadesh · Advances in transdisciplinary engineering · 2023
One of the most important elements in reducing cancer mortality rates is early identification. Oral cancer, for example, might be screened (much like squamous cell carcinoma.) by taking Papanicolaou test any dentist doctor. The only disadvantage is that manual processing of the vast amounts of data acquired by us is relatively expensive. CNN showed promise in differentiating among cancer and non-cancer cells which is allowing for fast automated cancer screening data processing. We look into different architectures that mainly focus on using texture information to classify cytological malignancy, driven by research which show chromatin texture is one of the most important discriminative variables for that reason. According to the findings, local binary patterns (lbps) in spire CNN classifiers outperform general-purpose CNN’S. This is true when varying different levels of data augmentation and pre-training are taken into account. We also created an interface for the application using the flask framework and deployed in to the AWS cloud for easy access from anywhere.