Graduate Research in Engineering and Technology (GRET)
Brain tumour detection is one of the most critical and arduous function in the domain of healthcare. Brain tumour, if not detected at an early stage, can be fatal. At the present time, detection and classification of brain tumour is done by the method of Biopsy which is very time-consuming and complex. . By looking at the brain MRI or CT scan, it is possible for the experts to identify whether tumour is present or not and the region of the tumour, but it is difficult to identify the small dissimilarities in the structure of tumour and classify it into types. Hence this manual process gets stuck here for verification of type of tumour. For the sole purpose overcoming the above-mentioned gigantic hurdles we have pursued this research of multi-class brain tumour detection using deep learning. Our project will help doctors in quick decision-making regarding detection of the tumour and its type as well, and due to the early detection of the disease the treatment can be initiated at the right time, resulting in speedy recovery of the patient. We propose a deep learning model employing Convolutional Neural Network architecture which we have implemented using Keras and Tensorflow because it yields to a better performance than the traditional ones. In our research work, CNN gained an accuracy of 94.95%. Further, we have integrated our model with a web-app which we have built using Streamlit. Hence, users can provide their MRI scans via our web-app and get their medical results in a quick and efficient manner.
Dhurkunde, Mukul Graduate Engineer; Kadam, Nayan Graduate Engineer; Trivedi, Mohak Graduate Engineer; Maru, Sandeep Graduate Engineer; and Shirke, Prajakta Assistant Professor
"Multi-class Brain Tumor Detection using Convolutional Neural Network,"
Graduate Research in Engineering and Technology (GRET): Vol. 1:
9, Article 1.
Available at: https://www.interscience.in/gret/vol1/iss9/1