Classification of Spine Abnormalities using Deep Learning

Citation

Ahmad, Muhammad Shahrul Zaim and Ab Aziz, Nor Azlina and Lim, Heng Siong (2024) Classification of Spine Abnormalities using Deep Learning. Proceedings of International Exchange and Innovation Conference on Engineering & Sciences (IEICES), 10. pp. 998-1004. ISSN 2434-1436

[img] Text
2024_p998.pdf - Published Version
Restricted to Repository staff only

Download (714kB)

Abstract

The spine provides structural support for the body. Identifying spinal abnormalities is crucial for assessing spinal health. Deep learning has gained popularity due to better performance compared to the traditional machine learning approaches. This study investigates the performance of a deep learning model in classifying different types of spinal abnormalities, namely scoliosis, spondylosis, and lesions using x-ray images. The deep learning model is trained and evaluated using two public datasets. The images in the first dataset are labelled to three classes; scoliosis, spondylosis, and normal, while the images in the second dataset are labelled to normal and lesion. The model was trained and tested on four different classification tasks. The deep learning model achieved accuracy of 100% for two binary and one multiclass classification of images from the first dataset and 97.5% accuracy for binary classification of images from the second dataset. The classification model would be a beneficial tool for medical professionals to perform initial diagnosis.

Item Type: Article
Uncontrolled Keywords: : Deep learning; spine abnormalities
Subjects: Q Science > Q Science (General) > Q300-390 Cybernetics
Divisions: Faculty of Engineering and Technology (FET)
Depositing User: Ms Nurul Iqtiani Ahmad
Date Deposited: 13 Jan 2025 04:42
Last Modified: 13 Jan 2025 04:42
URII: http://shdl.mmu.edu.my/id/eprint/13319

Downloads

Downloads per month over past year

View ItemEdit (login required)