Detection of Brain Tumors with AI Techniques

Statistical data related to health shows that cancer disease is increasing. Brain cancer is one of the types of cancer that can cause death, and it negatively affects the quality of life of patients struggling with this disease. Early diagnosis has a significant effect on extending the life span of patient groups struggling with brain cancer and brain tumors. Traditional methods and techniques used for early diagnosis have problems, such as requiring excessive human resources and the inability to report analysis results quickly. Improving the issues encountered in traditional methods is possible by developing innovative systems that can be easily applied and gain fast results. The study aims to classify benign and malignant cells with AI algorithms and design a system that reports results to patients and doctors via a web page. During the research, we applied AI algorithms such as K-NN, SVM and Transfer Learning in the "Br35H Brain Tumor Detection 2020" dataset. As a result of the experiments, the most successful performance was achieved with the EfficientNetB6 Transfer Learning algorithm, with 98%, and the results were reported to the user via the developed web page without any problems. The study results are expected to shed light new research using AI in the detection of brain cancer.