Diagnosis and classification of brain tumors using neural networks

The aim of the project was to develop an automated brain MRI diagnosis system with normal and abnormal classes using artificial intelligence, particularly Deep Learning. We investigated the efficiency of the proposed method in classifying the human brain into normal and abnormal classes

The anticipated benefit

We worked with MRI brain image datasets and used various forms of neural networks such as ANN, DNN, and simple neural networks for tumor classification. The anticipated benefit is to demonstrate that the proposed AI technique is fast, user-friendly, non-invasive, and cost-effective, thus offering significant advantages in the diagnosis and classification of brain tumors.

Results

We have shown that the proposed method is effective in classifying the human brain into normal and abnormal classes and offers significant benefits in diagnosing brain tumors. Additionally, we have highlighted the need for further research to improve tumor detection procedures by exploring various forms of neural networks and using less annotated images to achieve good results. We have also identified relevant parameters for evaluating system performance, such as accuracy, time, specificity, and efficiency, and emphasized the importance of introducing an automated expert system that can contribute to early diagnosis and improved treatment planning.