TY - GEN
T1 - Brain Tumor Detection on Magnetic Resonance Imaging (MRI) Images Using Convolutional Neural Network (CNN)
AU - Indraswari, Rarasmaya
AU - Ardan, Indira Salsabila
AU - Arifin, Agus Zainal
AU - Tjahyanto, Aris
AU - Rakhmawati, Nur Aini
AU - Kusumawardani, Renny
N1 - Publisher Copyright:
© 2022 Institute of Advanced Engineering and Science (IAES).
PY - 2022
Y1 - 2022
N2 - Brain cancer is one of the most malignant types of cancer. Several computer-Aided diagnostic (CAD) systems have been created to help clinicians analyze MRI (magnetic resonance imaging) images and find brain tumors. In general, CAD systems use conventional methods to classify images, in which the results are very dependent on the quality of object segmentation in the image. Meanwhile, tumor objects have different shapes, colors, and sizes, so the difficulty level of segmentation is quite high. On the other hand, currently, deep learning algorithms have been developed which are generally able to provide high accuracy results. Therefore, in this study, a tumor detection method on brain MRI images based on Convolutional Neural Network (CNN) is proposed. CNN method is one of the deep learning algorithms that combine the processes of segmentation, feature extraction, and classification into one. The system performs this task totally on its own through a training phase. However, the performance of the CNN method highly depends on the sufficient amount of training data, which usually cant be fulfilled in the case of medical data. Therefore, we proposed a CNN architecture developed from Inception V3in which the training process utilizes the transfer learning method to overcome the small amount of training set. The experimental result shows that the proposed model gives high performance on a small-size dataset with accuracy, precision, recall, and F1-score of 84%, 91%, 76%, and 83%, respectively. Moreover, InceptionV3 gives high efficiency compared to the other non-lightweight network architecture, such as ResNet50V2, InceptionResNetV2, and VGG16.
AB - Brain cancer is one of the most malignant types of cancer. Several computer-Aided diagnostic (CAD) systems have been created to help clinicians analyze MRI (magnetic resonance imaging) images and find brain tumors. In general, CAD systems use conventional methods to classify images, in which the results are very dependent on the quality of object segmentation in the image. Meanwhile, tumor objects have different shapes, colors, and sizes, so the difficulty level of segmentation is quite high. On the other hand, currently, deep learning algorithms have been developed which are generally able to provide high accuracy results. Therefore, in this study, a tumor detection method on brain MRI images based on Convolutional Neural Network (CNN) is proposed. CNN method is one of the deep learning algorithms that combine the processes of segmentation, feature extraction, and classification into one. The system performs this task totally on its own through a training phase. However, the performance of the CNN method highly depends on the sufficient amount of training data, which usually cant be fulfilled in the case of medical data. Therefore, we proposed a CNN architecture developed from Inception V3in which the training process utilizes the transfer learning method to overcome the small amount of training set. The experimental result shows that the proposed model gives high performance on a small-size dataset with accuracy, precision, recall, and F1-score of 84%, 91%, 76%, and 83%, respectively. Moreover, InceptionV3 gives high efficiency compared to the other non-lightweight network architecture, such as ResNet50V2, InceptionResNetV2, and VGG16.
KW - Convolutional Neural Network (CNN)
KW - Inception V3
KW - Magnetic Resonance Imaging (MRI)
KW - brain tumor
KW - deep learning
KW - image classification
UR - http://www.scopus.com/inward/record.url?scp=85142753812&partnerID=8YFLogxK
U2 - 10.23919/EECSI56542.2022.9946622
DO - 10.23919/EECSI56542.2022.9946622
M3 - Conference contribution
AN - SCOPUS:85142753812
T3 - International Conference on Electrical Engineering, Computer Science and Informatics (EECSI)
SP - 367
EP - 373
BT - Proceedings - 9th International Conference on Electrical Engineering, Computer Science and Informatics, EECSI 2022
A2 - Facta, Mochammad
A2 - Syafrullah, Mohammad
A2 - Riyadi, Munawar Agus
A2 - Subroto, Imam Much Ibnu
A2 - Irawan, Irawan
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 9th International Conference on Electrical Engineering, Computer Science and Informatics, EECSI 2022
Y2 - 6 October 2022 through 7 October 2022
ER -