Sustainable development of Flexible Assertion on Multi-Modal Classification of Brain Tumours using Deep Learning

Abstract

In the field of medical science, classifying brain tumours is vital. In order to get an effective and proper treatment for the disease, accurate and finding type of the brain tumour is much essential in the case of brain tumour treatment. In addition to providing treatment for tumours as early as possible, it also helps in saving a life by allowing medication to be administered in due course. DL has developed into a fantastic tool for medical professionals and researchers to act quickly and decisively with tumour patients. In this paper, we suggest Sustainable development of flexible approach aimed at multi-model organization of brain tumours using the popular deep learning architecture ResNet-50. By leveraging the flexibility of ResNet-50, we aim to achieve improved accuracy and robustness in classifying brain tumours across a diverse range of datasets. Our approach integrates multiple ResNet-50 models, each specialized in identifying specific tumour types, enabling a comprehensive classification framework. Experimental findings show that our strategy is successful and more accurate than other approaches. In this paper we provide an interface that can be used to classify and label the tumours. We used Keras and Tensorflow to create a cutting-edge Convolutional Neural Network (CNN) architecture to categorise 3 kinds of growth or tumours namely - Meningioma, Gliomaand Pituitary using ResNet50 algorithm. It is estimated that this model has a maximum mean accuracy score of 98.88%.