Brain Tumor Detection Using Convolutional Neural Network

The extraction of tumor areas from images is challenging since brain tumors have a diverse range of appearances and share many characteristics with normal tissues. Further-more, handling a sizable amount of data manually requires a time-consuming effort. This study developed a model for detecting brain tumors from two-dimensional MRI scans using the data augmentation method, CNN model, and pre-trained model. The experimental effort utilized the Kaggle dataset of tumors in varying sizes, shapes, and intensities. Deep learning, one of the most advanced technical methods for classifying and detecting tumors, has long been used to remove abnormal tumor areas from the brain. Advanced Artificial Intelligence and Neural Network classification algorithms may be useful in the early detection of brain cancers. To find brain tumors, CNN models were built using the VGG 16, and parameters were chosen to train this model to assess the literature solutions. VGG16 is one of the most effective CNN models due to its simplicity. Furthermore, the study discovered a way to use MRI for quick, efficient, and precise diagnose brain cancer. Using the Faster CNN, VGG 16 was used as the leading network to create feature maps that were then classified into tumor regions. The accuracy of the predictions was used to evaluate performance. The proposed technique was tested on a dataset of 253 MRI brain scans, with 155 revealing tumors. In MRI scans of the brain, the technique could be able to spot malignancies. The accuracy of VGG 16 was 99.94%, CNN was 97%, RestNet50 was 45.75%, and InceptionV3 was 48.85%.

Перевод пока недоступен