| |
|
|
|
Refined Convolutional Neural Networks Automated System for Brain Masses Detection using CT/MRI Diagnostic Scans |
|
|
|
PP: 749-759 |
|
|
|
doi:10.18576/amis/180407
|
|
|
|
Author(s) |
|
|
|
Lobna M. Abou El-Maged,
Israa AlQaisi,
Ghada A. Khouqeer,
Ahmed Elgarayhi,
Mohammed Sallah,
|
|
|
|
Abstract |
|
|
| A brain mass/tumor is considered to be a very fatal illness, exhibiting a diverse array of impacts on individuals’ general well-being. A neoplasm with abnormal cell proliferation, typically located inside or close to the cerebral region, is commonly known as a brain mass. Brain masses can manifest as either benign or malignant neoplasms. Medical practitioners employ many diagnostic methods to ascertain the nature of a patient’s brain tumor, distinguishing between benign and malignant tumors. Radiology images are currently viewed most often using deep learning techniques. Imaging methods include CT, MRI, PET, and ultrasound. CT and MRI scans are the most popular imaging, each with advantages and disadvantages. This paper has created an automatic system for detecting brain masses using CT and MRI scans. This is because these two types of X-rays each have their own advantages, and a radiologist would benefit from this method. The input image is subjected to testing by the system. If the image is identified as a CT-scan image, it uses the recommended Convolutional Neural Network (CNN) architecture to carry out diagnosis. Based on the achieved accuracy, F1-score, precision, and recall values of 98.01%, 98%, 99.7%, and 98.84%, respectively, the CNN architecture has proven to function exceptionally well. Alternatively, Reset101, a pre-trained convolutional neural network, can be used to diagnose the image in question if it is an MRI scan. The test results give 99.8%, 99.9%, 99.2%, and 99.55% for accuracy, precision, recall, and F1-score, respectively. |
|
|
|
|
 |
|
