Brain Tumor Detection and Segmentation Using Deep Learning

Brain Tumor Detection and Segmentation Using Deep Learning

Mrs. Charu Bolia

Department of Electronics and Telecommunication

Smt. Kashibai Navle College of Engineering

Vadgaon, Pune, India

charu.bolia@gmail.com

Sampada Shashikant Pasarkar

Department of Electronics and Telecommunication

Smt. Kashibai Navle College of Engineering

Vadgaon, Pune, India

sampadapasarkar27@gmail.com

Siddhant Nitin Kudale

Department of Electronics and Telecommunication

Smt. Kashibai Navle College of Engineering

Vadgaon, Pune, India

siddhantkudale18@gmail.com

Yash Pravin Deotale

Department of Electronics and Telecommunication

Smt. Kashibai Navle College of Engineering

Vadgaon, Pune, India

yashdeotale222@gmail.com

Abstract – A brain tumor is characterized by the presence of an abnormal mass or collection of cells within the brain. Tumors can be categorized as either cancerous (malignant) or noncancerous (benign). As a prevalent and highly destructive condition, malignant brain tumors significantly shorten life expectancy if not identified and treated promptly. Classifying brain tumors is a crucial step following detection to develop an effective treatment strategy. Early diagnosis not only improves the chances of effective medical intervention but can also be life-saving. The dataset used in this study includes MRI images of brains with and without tumors. These images undergo a series of preprocessing steps, including techniques like filtering, blurring, and cropping. The processed data is then fed into a Convolutional Neural Network (CNN) model based on VGG-16 architecture to determine the presence or absence of a tumor. The model also segments MRI images to identify and isolate regions of interest for analyzing tumor detection. The classification and detection of brain tumors are provided as the final output.

Keywords: Brain tumor, MRI images, Deep Learning, Convolutional Neural Networks (CNN), Visual Geometry Group (VGG16).