TY  - JOUR
AU  - Dr. M Kiran Kumar
AU  - B Meghana
AU  - N Sri Divya
AU  - R Sreya
AU  - K Krishna Chaitanyarao
PY  - 2025
DA  - 2025/03/06
TI  - A Machine Learning – Based Classification And Prediction Technique For Brain Tumor
JO  - Global Journal of Engineering Innovations and Interdisciplinary Research
VL  - 5
IS  - 1
AB  - The project's main goal is to solve the time-consuming and expensive problems that come with using MRI
scans to classify brain cancers. Brain cancers need a lot of effort and may not always be economical to
diagnose manually, noting the requirement for an automated remedy. The main objective of the project
is to use artificial intelligence (AI) techniques to create an automatic diagnosis system. The goal of the
system is to offer a rapid and precise method for differentiating between patients with glioma brain
tumors and pituitary tumors. High-quality photos from medical devices may help in the diagnosis and
early detection of various diseases. But the diagnosing procedure could take a while, and gathering
and storing pictures like this can be expensive. Synthetic AI-driven automated diagnostics could play a
significant role in addressing the issues of time and expense. Pre-trained deep-learning models might
offer an efficient method for categorizing medical images. In this study, we suggest two pretrained
models like ResNext101_32×8d and VGG19 to categorize two kinds of brain tumors like pituitary and
glioma. The suggested models are utilized on a dataset that includes 1,800 MRI images consisting of
two categories of diagnoses. Single-image super-resolution (SISR) A method is utilized on the MRI
images to categorize and improve their fundamental characteristics, allowing the suggested models to
improve specific features of the MRI images and aid the training process of the prototypes. We developed
these models utilizing the PyTorch and TensorFlow frameworks with adjusting hyper-parameters and
enhancing data. The evaluation of the models is examined. utilizing different metrics, and the findings
show that the elevated testing accuracies and minimal loss fees for both models. The suggested models,
when utilized on MRI images, aim to deliver a swift and precise method for differentiating between
patients with pituitary and glioma tumors. These models may assist physicians and radiologists in the
evaluation of patients with brain tumors.
SN  - 3066-1226
UR  - https://dx.doi.org/10.33425/3066-1226.1077
DO  - 10.33425/3066-1226.1077