Histopathologic Oral Cancer Prediction Using Oral Squamous Cell Carcinoma Biopsy Empowered with Transfer Learning

Abstract

Oral cancer is a dangerous and extensive cancer with a high death ratio. Oral cancer is the most usual cancer in the world, with more than 300,335 deaths every year. The cancerous tumor appears in the neck, oral glands, face, and mouth. To overcome this dangerous cancer, there are many ways to detect like a biopsy, in which small chunks of tissues are taken from the mouth and tested under a secure and hygienic microscope. However, microscope results of tissues to detect oral cancer are not up to the mark, a microscope cannot easily identify the cancerous cells and normal cells. Detection of cancerous cells using microscopic biopsy images helps in allaying and predicting the issues and gives better results if biologically approaches apply accurately for the prediction of cancerous cells, but during the physical examinations microscopic biopsy images for cancer detection there are major chances for human error and mistake. So, with the development of technology deep learning algorithms plays a major role in medical image diagnosing. Deep learning algorithms are efficiently developed to predict breast cancer, oral cancer, lung cancer, or any other type of medical image. In this study, the proposed model of transfer learning model using AlexNet in the convolutional neural network to extract rank features from oral squamous cell carcinoma (OSCC) biopsy images to train the model. Simulation results have shown that the proposed model achieved higher classification accuracy 97.66% and 90.06% of training and testing, respectively.

IoMT-Based Osteosarcoma Cancer Detection in Histopathology Images Using Transfer Learning Empowered with Blockchain, Fog Computing, and Edge Computing

“…Researchers have developed automatic detection systems for analyzing and classifying medical images (X-ray, ultrasound imaging, MRI, CT scan, histology images, etc.) to detect various types of malignancies and tumors as technology advances [ 11 , 12 , 13 , 14 , 15 , 16 , 17 ]. Several automated systems for categorizing medical images use handcrafted features [ 18 , 19 , 20 , 21 ], whereas other systems use DLMs to extract and classify medical images [ 22 , 23 , 24 , 25 , 26 ].…”

“…The prior research has the following flaws: It used handcrafted features for tumor classification, which is a tedious task [ 17 , 18 , 19 , 20 ]; MRI and CT scans were used for classification, which cannot provide cellular information [ 34 , 35 , 36 ]; It reported low accuracy [ 38 , 40 , 43 ]; There was no mechanism to ensure the security and privacy of patient data [ 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 ]. …”

“…It used handcrafted features for tumor classification, which is a tedious task [ 17 , 18 , 19 , 20 ];…”

Assessment of Information Extraction Techniques, Models and Systems

“…Recent comparative study [72] shows that GROBID, CERMINE and ParsCit pose best results among the various open-source extractors. In future, the schemes involving deep learning models and other hybrid intelligent approaches such as federated learning and transfer learning [91][92][93][94][95][96][97][98][99][100] can be anticipated more successful and worthy to explore for the IE from published scholarly articles of diverse nature.…”

IoMT-Based Mitochondrial and Multifactorial Genetic Inheritance Disorder Prediction Using Machine Learning

“…the accuracy, miss-classification rate, sensitivity, precision, and F1 score values are calculated by using the formulas mentioned below[37,[40][41][42][43][44][45][46][47][48][49][50][51].…”