Archive of SID
Multidisciplinary Cancer Investigation Original Article
© 2020. Multidisciplinary Cancer Investigation
DOI:10.30699/acadpub.mci.4.1.5
Submitted: 9 November 2019
Revised: 12 December 2019
Accepted: 23 December 2019
e-Published: 1 January 2020
Keywords:
Brain Neoplasms
Nerve Net
Magnetic Resonance Imaging
Introduction: Brain tumors such as glioma are among the most aggressive lesions,
which result in a very short life expectancy in patients. Image segmentation is highly
essential in medical image analysis with applications, particularly in clinical practices
to treat brain tumors. Accurate segmentation of magnetic resonance data is crucial for
diagnostic purposes, planning surgical treatments, and also follow-up evaluation. Man-
ual segmentation of a large volume of MRI data is a time-consuming endeavor, and this
necessitates employing automatic segmentation techniques that are both accurate and
reliable. However, the vast spatial and structural diversity of brain tissue poses serious
challenges for this procedure. The current study proposed an automatic segmentation
method based on convolutional neural networks (CNN), where weights of a pre-trained
network were used as initial weights of neurons to prevent possible overftting in the
training phase.
Methods: As tumors were diverse in their shape, size, location, and overlapping with
other tissue, it was decided to exploit a fexible and extremely effcient architecture
tailored to glioblastoma. To remove some of the overlapping diffculties, morphological
operators as a pre-processing step were utilized to strip the skull.
Results: The proposed CNN had a hierarchical architecture to exploit local and global
contextual features to handle both high- and low-grade glioblastoma. To address bias-
ing stem from the imbalance of tumor labels, dropout was employed and a stochastic
pooling layer was proposed.
Conclusions: Experimental results reported on a dataset of 400 brain MR images sug-
gested that the proposed method outperformed the currently published state-of-the-art
approach in terms of various image quality assessment metrics and achieved magnitude
fold speed-up.
A Hierarchical Convolutional Neural Network Architecture
for Brain Tumor Segmentation in 3D Brain Magnetic
Resonance Imaging
Ayoub Adineh-vand
1
, Gholamreza Karimi
1, *
, Mozafar Khazaei
2
1
Electrical Engineering Department, Faculty of Engineering, Razi University,
Kermanshah, Iran
2
Fertility and Infertility Research Center, Health Technology Institute, Kermanshah
University of Medical Sciences, Kermanshah, Iran
*Corresponding author: Gholamreza Karimi, Electrical Engineering, Faculty
of Engineering, Razi University, Postal Code: 67149, Kermanshah, Iran. Tel:
+988334343218, Fax: +988334343211, E-mail: ghkarimi@razi.ac.ir
Image segmentation aims at partitioning images
into homogeneous regions by employing spatial-
spectral attributes of the image. A segmentation
algorithm assigns a unique label to each pixel
and defnes segmented regions based on all pixels
serving certain criteria. This task has an important
INTRODUCTION
January 2020, Volume 4, Issue 1
www.SID.ir