Deep Learning Based Brain Tumor Detection Using VGG16 Model

Computer Science and Engineering

In the realm of brain tumor medical image processing, the segmentation of brain tumors is a vital and demanding undertaking. Relying on human-assisted manual classification for this task can lead to inaccurate predictions and diagnoses. Moreover, the difficulty is exacerbated when there is a substantial amount of data that requires assistance. In this study, we present a deep learning approach utilizing the VGG16 architecture to process 2D Magnetic Resonance brain images (MRI) and accurately distinguish between normal and abnormal cases based on texture and statistical features. By employing transfer learning techniques, we anticipate achieving an accuracy of 90\% or higher. The primary objective of image segmentation in medical image processing is the detection of tumors or lesions. Enhancing the sensitivity and specificity of tumor or lesion identification has become a central challenge in medical imaging with the aid of Computer-Aided Diagnostic (CAD) systems. However, manual segmentation of tumors or lesions is a time-consuming, challenging, and burdensome task, particularly due to the large number of MRI images generated in routine medical practice. To address this, we propose an efficient and proficient method that facilitates brain tumor segmentation and detection without the need for human assistance, leveraging the VGG16 architecture and transfer learning techniques.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.