A Sequential Machine Learning-cum-Attention Mechanism for Effective Segmentation of Brain Tumor

Tahir Mohammad Ali; Ali Nawaz; Attique Ur Rehman; Rana Zeeshan Ahmad; Abdul Rehman Javed; Thippa Reddy Gadekallu; Chin-Ling Chen; Chih-Ming Wu

doi:10.3389/fonc.2022.873268

A Sequential Machine Learning-cum-Attention Mechanism for Effective Segmentation of Brain Tumor

Front Oncol. 2022 Jun 1:12:873268. doi: 10.3389/fonc.2022.873268. eCollection 2022.

Authors

Tahir Mohammad Ali¹, Ali Nawaz¹, Attique Ur Rehman^{1

2}, Rana Zeeshan Ahmad³, Abdul Rehman Javed⁴, Thippa Reddy Gadekallu⁵, Chin-Ling Chen^{6

7

8}, Chih-Ming Wu⁹

Affiliations

¹ Department of Computer Science, GULF University for Science and Technology, Mishref, Kuwait.
² Department of Software Engineering, University of Sialkot, Sialkot, Pakistan.
³ Department of Information Technology, University of Sialkot, Sialkot, Pakistan.
⁴ Department of Cyber Security, Air University, Islamabad, Pakistan.
⁵ School of Information Technology and Engineering, Vellore Institute of Technology, Vellore, India.
⁶ School of Information Engineering, Changchun Sci-Tech University, Changchun, China.
⁷ School of Computer and Information Engineering, Xiamen University of Technology, Xiamen, China.
⁸ Department of Computer Science and Information Engineering, Chaoyang University of Technology, Taichung, Taiwan.
⁹ School of Civil Engineering and Architecture, Xiamen University of Technology, Xiamen, China.

Abstract

Magnetic resonance imaging is the most generally utilized imaging methodology that permits radiologists to look inside the cerebrum using radio waves and magnets for tumor identification. However, it is tedious and complex to identify the tumorous and nontumorous regions due to the complexity in the tumorous region. Therefore, reliable and automatic segmentation and prediction are necessary for the segmentation of brain tumors. This paper proposes a reliable and efficient neural network variant, i.e., an attention-based convolutional neural network for brain tumor segmentation. Specifically, an encoder part of the UNET is a pre-trained VGG19 network followed by the adjacent decoder parts with an attention gate for segmentation noise induction and a denoising mechanism for avoiding overfitting. The dataset we are using for segmentation is BRATS'20, which comprises four different MRI modalities and one target mask file. The abovementioned algorithm resulted in a dice similarity coefficient of 0.83, 0.86, and 0.90 for enhancing, core, and whole tumors, respectively.

Keywords: BRATS; MRI; UNET; VGG19; attention mechanism; brain tumor segmentation.