TY - GEN
T1 - The Classification of Brain Tumours by Means of Feature-Based Transfer Learning
AU - Wu, Chengzhangzheng
AU - Yang, Junqing
AU - Liu, Taimingwang
AU - Tan, Andrew
AU - Luo, Yang
AU - Razman, Mohd Azraai Mohd
AU - Majeed, Anwar P.P.Abdul
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.
PY - 2024
Y1 - 2024
N2 - Brain tumours are abnormal growths of cells in the brain and can be life-threatening if not detected early. Traditionally, radiologists manually assess magnetic resonance imaging (MRI) scans of the brain to identify and evaluate brain tumours; however, this process is prone to misinterpretation. This study investigates the application of a deep learning technique known as feature-based transfer learning to automate brain tumour detection from MRI images. A dataset of MRI scans labelled with different types of brain tumours was utilised in the study, in which a MobileNet pre-trained convolutional neural network was used to extract discriminative features from the images. The different classes of the tumors were then classified by three vanilla machine learning models, i.e., kNearest Neighbors (kNN), Support Vector Machine (SVM) and Logistic Regression (LR). The study showed that the MobileNet + LR pipeline could distinguish the classes well. The proposed method demonstrates its potential for augmenting and enhancing radiologist assessment of medical imaging.
AB - Brain tumours are abnormal growths of cells in the brain and can be life-threatening if not detected early. Traditionally, radiologists manually assess magnetic resonance imaging (MRI) scans of the brain to identify and evaluate brain tumours; however, this process is prone to misinterpretation. This study investigates the application of a deep learning technique known as feature-based transfer learning to automate brain tumour detection from MRI images. A dataset of MRI scans labelled with different types of brain tumours was utilised in the study, in which a MobileNet pre-trained convolutional neural network was used to extract discriminative features from the images. The different classes of the tumors were then classified by three vanilla machine learning models, i.e., kNearest Neighbors (kNN), Support Vector Machine (SVM) and Logistic Regression (LR). The study showed that the MobileNet + LR pipeline could distinguish the classes well. The proposed method demonstrates its potential for augmenting and enhancing radiologist assessment of medical imaging.
KW - Brain Tumor
KW - Computer Aided Diagnosis
KW - Deep Learning
KW - Feature-based Transfer Learning
KW - Machine Learning
UR - http://www.scopus.com/inward/record.url?scp=85211371110&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-70687-5_13
DO - 10.1007/978-3-031-70687-5_13
M3 - Conference Proceeding
AN - SCOPUS:85211371110
SN - 9783031706868
T3 - Lecture Notes in Networks and Systems
SP - 123
EP - 128
BT - Robot Intelligence Technology and Applications 8 - Results from the 11th International Conference on Robot Intelligence Technology and Applications
A2 - Abdul Majeed, Anwar P. P.
A2 - Yap, Eng Hwa
A2 - Liu, Pengcheng
A2 - Huang, Xiaowei
A2 - Nguyen, Anh
A2 - Chen, Wei
A2 - Kim, Ue-Hwan
PB - Springer Science and Business Media Deutschland GmbH
T2 - 11th International Conference on Robot Intelligence Technology and Applications, RiTA 2023
Y2 - 6 December 2023 through 8 December 2023
ER -