SpineResUnet: Classification and Prediction of Spinal Tuberculosis Disease on Exploiting the Structural and Texture Dependencies

SpineResUnet: Classification and Prediction of Spinal Tuberculosis Disease on Exploiting the Structural and Texture Dependencies
  IJETT-book-cover           
  
© 2023 by IJETT Journal
Volume-71 Issue-10
Year of Publication : 2023
Author : Askarali K.T, E.J. Thomson Fredrik
DOI : 10.14445/22315381/IJETT-V71I10P214

How to Cite?

Askarali K.T, E.J. Thomson Fredrik, "SpineResUnet: Classification and Prediction of Spinal Tuberculosis Disease on Exploiting the Structural and Texture Dependencies," International Journal of Engineering Trends and Technology, vol. 71, no. 10, pp. 156-162, 2023. Crossref, https://doi.org/10.14445/22315381/IJETT-V71I10P214

Abstract
Spinal Tuberculosis is a most important and very dangerous extrapulmonary disease type of skeletal tuberculosis, which causes destruction, collapse of vertebrae and angulations of the vertebral column. Hence, detecting the disease early is becoming important to prevent serious complications such as spinal deformity and permanent neurological deficit. Many machine learning and deep learning models have been presented to detect and classify the disease. However, those techniques lead to various challenges in appearance and are complex in fully exploiting the dependency between the structural and texture characteristics of the image. In order to handle the challenges, a new deep learning framework named SpineResUnet was to achieve the classification of spinal tuberculosis disease on volumetric MR Images. SpineGraphNet is composed of automated segmentation and classification of the image. Initially, segmentation of the collapsed region of vertebrate or spinal deformity region is carried out using graph convolution segmentation network as fine-grained segments. The segmented region of the image is processed further towards classifying the segmented region into classes using an adjacency matrix containing semantic features in ResNet and U-Net with Skip Connection. The classes represent kyphosis, gibbus formation and osteomyelitis of the spinal tuberculosis. The proposed model successfully captures the dependency implicitly and explicitly. Experimental analysis of T2 weighted volumetric MR image of 100 subjects has been collected from the KMCH hospital, Coimbatore. In this, 60% of images have been employed to train corresponding experts in Spinal Tuberculosis disease characterization, 20% has been used as testing, and the remaining 20% of the image has been taken for 5-fold cross-validation through a confusion matrix. The proposed model exhibited optimal performance in terms of dice coefficient, specificity, and sensitivity compared against conventional approaches.

Keywords
Deep learning, Gibbus formation, Kyphosis, Osteomyelitis, Spinal tuberculosis.

References
[1] Cheng Chen et al., “Localization and Segmentation of 3D Intervertebral Discs in MR Images by Data Driven Estimation,” IEEE Transactions on Medical Imaging, vol. 34, no. 8, pp. 1719–1729, 2015.
[CrossRef] [Google Scholar] [Publisher Link]
[2] Xiaomeng Li et al., “3D Multi-Scale FCN with Random Modality Voxel Dropout Learning for Intervertebral Disc Localization and Segmentation from Multi-Modality MR Images,” Medical Image Analysis, vol. 45, pp. 41–54, 2018.
[CrossRef] [Google Scholar] [Publisher Link]
[3] Nikolas Lessmann et al., “Iterative Fully Convolutional Neural Networks for Automatic Vertebra Segmentation and Identification,” Medical Image Analysis, vol. 53, pp. 142–155, 2019.
[CrossRef] [Google Scholar] [Publisher Link]
[4] Shuo Li et al., “MRLN: Multi-Task Relational Learning Network for MRI Vertebral Localization, Identification and Segmentation,” IEEE Journal of Biomedical and Health Informatics, vol. 24, no. 10, pp. 2902-2911, 2020.
[CrossRef] [Google Scholar] [Publisher Link]
[5] Faezeh Fallah et al., “Simultaneous Volumetric Segmentation of Vertebral Bodies and Intervertebral Discs on Fat Water MR Images,” IEEE Journal of Biomedical and Health Informatics, vol. 23, no. 4, pp. 1692–1701, 2019.
[CrossRef] [Google Scholar] [Publisher Link]
[6] Sepp Hochreiter, and Jürgen Schmidhuber, “Long Short-Term Memory,” Neural Computation, vol. 9, no. 8, pp. 1735-1780, 1997.
[CrossRef] [Google Scholar] [Publisher Link]
[7] Zhongyi Han et al., “Spine-GAN: Semantic Segmentation of Multiple Spinal Structures,” Medical Image Analysis, vol. 50, pp. 23–35, 2018.
[CrossRef] [Google Scholar] [Publisher Link]
[8] Tianyang Li et al., “S3egAnet: 3D Spinal Structures Segmentation via Adversarial Nets,” IEEE Access, vol. 8, pp. 1892–1901, 2019.
[CrossRef] [Google Scholar] [Publisher Link]
[9] Hao Chen et al., “3D Fully Convolutional Networks for Intervertebral Disc Localization and Segmentation,” International Conference on Medical Imaging and Augmented Reality, pp. 375–382, 2016.
[CrossRef] [Google Scholar] [Publisher Link]
[10] Anjany Sekuboyina, “Attention-Driven Deep Learning for Pathological Spine Segmentation,” International Workshop on Computational Methods and Clinical Applications in Musculoskeletal Imaging, pp. 108–119, 2017.
[CrossRef] [Google Scholar] [Publisher Link]
[11] Rens Janssens, Guodong Zeng, and Guoyan Zheng, “Fully Automatic Segmentation of Lumbar Vertebrae from CT Images Using Cascaded 3D Fully Convolutional Networks,” IEEE 15th International Symposium on Biomedical Imaging, pp. 893–897, 2018.
[CrossRef] [Google Scholar] [Publisher Link]
[12] Chuanbo Wang et al., “Fully Automatic Intervertebral Disc Segmentation using Multimodal 3D U-Net,” 2019 IEEE 43rd Annual Computer Software and Applications Conference, vol. 1, pp. 730–739, 2019.
[CrossRef] [Google Scholar] [Publisher Link]
[13] Nassim Guerroumi et al., “Automatic Segmentation of the Scoliotic Spine from Mr Images,” 2019 IEEE 16th International Symposium on Biomedical Imaging, pp. 480–484, 2019.
[CrossRef] [Google Scholar] [Publisher Link]
[14] Jayashree Shedbalkar, K. Prabhushetty, and R.H. Havaldar, “UNet and Transformer-Based Model for Multi-Modality Brain Tumor Segmentation,” SSRG International Journal of Electronics and Communication Engineering, vol. 10, no. 8, pp. 22-35, 2023.
[CrossRef]  [Publisher Link]
[15] Olaf Ronneberger, Philipp Fischer, and Thomas Brox, “U-Net: Convolutional Networks for Biomedical Image Segmentation,” International Conference on Medical Image Computing and Computer-Assisted Intervention, pp. 234–241, 2015.
[CrossRef] [Google Scholar] [Publisher Link]
[16] Liang-Chieh Chen et al., “Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation,” Proceedings of the European Conference on Computer Vision, 2018, pp. 801–818, 2018.
[Google Scholar] [Publisher Link]
[17] Ke Gong et al., “Graphonomy: Universal Human Parsing via Graph Transfer Learning,” Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 7450–7459, 2019.
[Google Scholar] [Publisher Link]
[18] Anjany Sekuboyina et al., “A Localisation-Segmentation Approach for Multi-Label Annotation of Lumbar Vertebrae using Deep Nets,” arXiv, 2017.
[CrossRef] [Google Scholar] [Publisher Link]
[19] Thomas N. Kipf, and Max Welling, “Semi-Supervised Classification with Graph Convolutional Networks,” arXiv, 2016.
[CrossRef] [Google Scholar] [Publisher Link]
[20] M. Nisha et al., “Automatic Hippocampus Segmentation Model for MRI of Human Head through Semi-Supervised Generative Adversarial Networks,” Journal Neuro Quantology, vol. 20, no. 6, 2022.
[Google Scholar] 
[21] Kaiming He et al., “Deep Residual Learning for Image Recognition,” In 2016 IEEE Conference on Computer Vision and Pattern Recognition, pp. 770-778, 2016.
[Google Scholar] [Publisher Link]