TY - GEN

T1 - Brain Tumor Classification based on MR Images using GAN as a Pre-Trained Model

AU - Yerukalareddy, Dinesh Reddy

AU - Pavlovskiy, Evgeniy

N1 - Funding Information: The reported study was funded by RFBR according to the research project No 19-29-01103. Publisher Copyright: © 2021 IEEE.

PY - 2021/5/26

Y1 - 2021/5/26

N2 - In the medical industry, misdiagnosis of disease is acknowledged as the most common and harmful medical errors as it can cost a human life. Radiologists require a lot of time to manually annotate and segment the images. Over the several years, deep learning has been playing a vital role in the field of computer vision. One of its key uses in the medical industry is to minimize misdiagnosis and the amount of time taken to annotate and segment the images. In this paper, a new deep learning approach for brain tumor classification on MRI Images is introduced. A deep neural network is pretrained as a discriminator in a generative adversarial network (GAN) on MR Images by using multi-scale gradient GAN (MSGGAN) with auxiliary classification to extract the features and to classify the images. In the discriminator, one of the fully connected blocks acts as an auxiliary classifier and the other fully connected block acts as an adversarial. The fully connected layers of the auxiliary block are fine-tuned to classify the type of tumor. The proposed approach is tested on two publicly available MRI datasets as a whole, consists of four types of brain tumors (glioma, meningioma, pituitary, and no tumor). Our proposed method achieved 98.57% accuracy which is better as compared to state of art methods. Also, our method appears to be a useful technique when the availability of medical images is limited.

AB - In the medical industry, misdiagnosis of disease is acknowledged as the most common and harmful medical errors as it can cost a human life. Radiologists require a lot of time to manually annotate and segment the images. Over the several years, deep learning has been playing a vital role in the field of computer vision. One of its key uses in the medical industry is to minimize misdiagnosis and the amount of time taken to annotate and segment the images. In this paper, a new deep learning approach for brain tumor classification on MRI Images is introduced. A deep neural network is pretrained as a discriminator in a generative adversarial network (GAN) on MR Images by using multi-scale gradient GAN (MSGGAN) with auxiliary classification to extract the features and to classify the images. In the discriminator, one of the fully connected blocks acts as an auxiliary classifier and the other fully connected block acts as an adversarial. The fully connected layers of the auxiliary block are fine-tuned to classify the type of tumor. The proposed approach is tested on two publicly available MRI datasets as a whole, consists of four types of brain tumors (glioma, meningioma, pituitary, and no tumor). Our proposed method achieved 98.57% accuracy which is better as compared to state of art methods. Also, our method appears to be a useful technique when the availability of medical images is limited.

KW - ACGAN

KW - Deep Learning

KW - GAN

KW - MRI

KW - MSG-GAN

KW - Neural Network

UR - http://www.scopus.com/inward/record.url?scp=85112372790&partnerID=8YFLogxK

U2 - 10.1109/CSGB53040.2021.9496036

DO - 10.1109/CSGB53040.2021.9496036

M3 - Conference contribution

AN - SCOPUS:85112372790

SN - 978-1-6654-3150-7

T3 - Proceedings - 2021 IEEE Ural-Siberian Conference on Computational Technologies in Cognitive Science, Genomics and Biomedicine, CSGB 2021

SP - 380

EP - 384

BT - Proceedings - 2021 IEEE Ural-Siberian Conference on Computational Technologies in Cognitive Science, Genomics and Biomedicine, CSGB 2021

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2021 IEEE Ural-Siberian Conference on Computational Technologies in Cognitive Science, Genomics and Biomedicine, CSGB 2021

Y2 - 26 May 2021 through 28 May 2021

ER -