Paper Key : IRJ************481
Author: Cuba A R,Aarthika S,Gayatri A,Sowmiya S R
Date Published: 13 Apr 2025
Abstract
Brain diseases classification is the most challenging process due to their sensitivity, the difficulty of executing operations, and the high expenses. Brain tumors are a significant and potentially life-threatening medical condition that demands accurate and timely diagnosis for effective treatment planning. Magnetic Resonance Imaging (MRI) has emerged as a vital tool in the field of medical imaging, offering detailed and non-invasive visualization of the brain's internal structures. Brain tumor detection using MRI images has become a cornerstone in the diagnosis and management of these conditions. The proposed approach begins by pre- processing the dataset, which consists of MRI scans and clinical data from individuals with different brain disease conditions. Dataset consists a diverse range of brain images, including both tumor and non-tumor cases. The dataset is divided into training and testing sets, ensuring a balanced distribution of samples for each class. On MRI scans, many procedures are needed to identify tumors, including image pre-processing, feature extraction, and classification. Convolutional Neural Networks with Long Short-Term Memory (LSTM) will be using in this project. The hybrid model, known for its deep architecture and robust feature extraction capabilities, it will be employed alongside the LSTM, a highly efficient architecture designed for image classification tasks. The findings will be revealing several notable advantages of LSTM model. The hybrid learning models automatically learning hierarchical features from raw image data, eliminating the need for manual feature engineering. This feature extraction capability enables CNNs to capture subtle and complex patterns within brain images, enhancing their diagnostic accuracy. The results of proposed work will demonstrate the effectiveness of the proposed method to detect the brain tumors, which will achieve high accuracy, precision, recall and F1-measure. Based on model accuracy, user can input the test brain MRI image to predict the tumor with types and also provide the diagnosis details about predicted disease. Experimental results shows that the proposed model provide improved efficiency in disease prediction.
DOI Requested