← Back to Publications List

Automated Detection and Classification of Brain Tumors Using Image Processing Techniques

Students & Supervisors

Student Authors
Tithi Kirttania
Bachelor of Science in Computer Science & Engineering, FST
Liyan Ghosh
Bachelor of Science in Computer Science & Engineering, FST
Md.masidur Rahman Redoy
Bachelor of Science in Computer Science & Engineering, FST
Supervisors
Md.mortuza Ahmmed
Associate Professor, Faculty, FST

Abstract

Brain tumors pose a significant threat to human life because they disrupt the operation of body components in human individuals. Based on Cancer.Net, it is estimated that 24,810 adults in the United States will be diagnosed with primary malignant brain and spinal cord tumors in 2023. This figure comprises 14,280 men and 10,530 women. But in complicated cases like meningiomas, even for doctors, to define the exact location of the tumor could be difficult. The aim of this study is to devise a new technique for brain tumor detection and classification based on the capabilities of deep learning models and image processing algorithms. A longitudinal dataset over the time frame 2000-2024 was built out of the official data in the retrospective analysis, based on reported data from cancer.NET. The dataset includes variables like detections counts,clinical trial counts,imaging modalities(CT,MRI.MRI+PET) and significant events like introduction of CNNs,transfer learning. The dataset was preprocessed using image processing to reap relevant features for model training.The outcome of the dataset was highly encouraging, as the CNN model was able to detect and classify with an accuracy rate of 93%.The multimodal imaging model was even more accurate at 92%, and it surpassed the InceptionV3 model, which was found to be accurate at a rate of 91%.The high accuracy rates of the models establish their potential for detecting and classifying brain tumors based on MRI+PET images. These results show the capability of deep learning models with the ability to enable accurate brain tumor evaluating with enhanced medical image analysis scope.The paper proposes a method that suits three pre-trained models for brain tumor classification, including CNN, VGG16, and Cloud-based diagnosis. The purpose of the paper is to identify and classify brain tumors from MRI scans.

Keywords

Tumor Detection Image Processing Convolutional Neural Networks (CNN) Transfer Learning Brain Tumor Automated Diagnosis MRI PET.

Publication Details

  • Type of Publication:
  • Conference Name: BURS 1st National Youth Research Summit 2025
  • Date of Conference: 18/10/2025 - 18/10/2025
  • Venue: University of Barishal
  • Organizer: Barishal University Research Society (BURS)
View on Publisher