IDENTIFYING AND CLASSIFYING ORAL CANCER BASED ON DEEP TRANSFER LEARNING

Computer Science and Engineering

The "head and neck cancers" that occur most often worldwide are oral cancers. Oral cancer is a common, intricate, and very dangerous type of cancer. In India, oral malignant growth positions seventh among all tumors, with 130,000 fatalities annually. The tumor has an impact on the tonsils, salivary glands, face, mouth, and neck. The majority of oral cancer cases are only identified in advanced stages due to an absence of public mindfulness. In this specific circumstance, computerized reasoning (artificial intelligence) and AI (ML) models are utilized since it is critical to distinguish sicknesses from the beginning for improved results. The ongoing work presents the Oral Disease recognition and Characterization Model (AIDTL-OCCM), which is controlled by man-made reasoning and profound exchange learning. The proposed AIDTL-OCCM model's main objective is to identify oral cancer utilizing AI and image processing methods. The AIDTL-OCCM model under consideration uses a fuzzy-based contrast enhancing method. Then, a suitable set of deep features is produced using the densely-connected networks (DenseNet-169) model. Additionally, the Autoencoder (AE) model of the Chimp Optimization Algorithm (COA) is employed to identify and classify oral cancer. To be able to choose the best AE model parameters, COA is also used. Benchmark datasets were used as the basis for a wide range of experimental investigations, and the outcomes were examined from a number of angles. With a maximum accuracy of 90.08%, the results of the thorough experimental investigation demonstrated the AIDTL-OCCM model's superior performance over other methods.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.