Image Classification & Detection of Brain Tumor

Accurate and early detection of brain tumors is very important for planning effective treatments and improving patients' outcomes. This paper presents an CNN-based automated recognition of brain MRI images as a classifier between two possibilities: tumor-positive (“Yes”) or tumor-negative (“No”). The approach considered here is a bit different from ones which focus on the differentiation of specific types of tumour; it only emphasises that there is a tumour, which would make it as simple and efficient as possible. The CNN model for brain tumor detection is trained over a set of labeled MRI images. Monitoring all the metrics in terms of accuracy, validation accuracy, and loss carefully can provide further insight about how the model is learning and performing as a whole. The distinguishing capability between tumor and non-tumor images with high precision points out its feasibility as a reliable tool for diagnostics. In order to make the CNN model more user-friendly and very simple to use, a Flaskbased web application has been developed. This Web interface helps medical professionals and any other users upload MRI images and gives them real-time results that indicate the absence or presence of a tumor. It is user-friendly, thereby making it quicker to make decisions and preventing manual errors in interpretation. The experimental results show the accuracy and validation loss at a pretty low rate, which really shows that the model is performing well in tumour detection. It portrays Artificial intelligence and deep learning at its best potential in the examination of medical pictures and helps with the development of scalable, accurate, and efficient solutions that can easily detects brain tumor. Its focus is only on the binary classification and addresses one of the most important diagnostic needs, hence forming a valuable tool in improving early diagnosis and treatment.