RESUMO
The Clinical Oncology of American Society report in 2016 predicted deaths are increased upto 9570 due to oral cancer. This cancer occurs due to abnormal tissue growth in the oral cavity. This cancer has limited symptoms, so, it has been difficult to recognize in the early stages. To reduce the death rate of this oral cavity cancer, an automatic system has been developed by applying the optimization techniques in both image processing and machine learning techniques. Even though these methods are successfully recognizing the cancer, the detection accuracy is still one of the major issues because of complex oral tissue structure. So, this paper introduces the Gravitational Search Optimized Echo state neural networks for predicting the oral cancer with effective manner. Initially the X-ray images are collected from the oral cancer database which contains several noises that has to be eliminated with the help of the adaptive wiener filter. Then the affected part has been segmented with the help of the enhanced Markov Stimulated Annealing and the features are derived from segmented region. The derived features are analyzed with the help of the proposed classifier. The excellence of the oral cancer detection system is evaluated using simulation results.
