Multidomain Feature Level Fusion for Classification of Lumbar Intervertebral Disc Using Spine MR Images

Abstract

Grading of discs is essential for the assessment of degeneration progression which subsequently plays a vital role in decision making in the removal of a disc. In particular, Pfirrmann’s five-scale (1–5) scoring system is widely used in MR image modality for grading the discs. In this study, we have presented a contemporary semiautomatic feature level fusion approach for the classification of inter-vertebral discs. The data of T2-weighted lumbar MR scans in sagittal plane were collected from 120 distinct subjects. In total, 1123 inter-vertebral disc images were obtained upon performing image augmentation. The experts have segregated the discs into five categories as per Pfirrmann’s criteria. This segregation is utilized as ground truth label data for classification. Furthermore, two feature extraction techniques are exploited, one from spatial domain and other follows deep learning process. A popular Local Binary Pattern (LBP) texture descriptor extracts features from spatial domain. In addition, a popular pre-trained Convolution Neural Network (CNN), which acts as a feature extractor, extracts deep features. The training procedure using SVM classifier yields a model built from post-fusion feature vectors. Furthermore, to estimate the model’s performance, a 5-fold cross-validation is performed by computing principal component analysis as well as without dimensionality reduction. Experiment results obtained on our dataset indicate that after dimensionality reduction, SVM classifier with various kernel functions yields the accuracy up to 92%. A quantitative analysis of the classifier model is presented for parameters, namely – Accuracy, Area Under Curve (AUC), Specificity, Sensitivity, and F1 measure.

KEYWORDS:

• Classification
• Deep learning
• Intervertebral disc degeneration
• MRI images
• SVM

Acknowledgements

The authors are thankful to Dr Hemant Borse, MD, a Consultant Radiologist, Samarth Diagnostic center Nasik, India and Dr Rajesh Jawale, MD, a Consultant Radiologist, Wockhardt Hospital Nasik (M.S.) India who have provided very deep insights and medical expertise that greatly assisted this work. The authors are incalculably grateful to Dr. Hemant Borse and his team for allowing access to spine MR image dataset for this research work. Dataset were pseudonymised and all identifying data were omitted.

Additional information

Notes on contributors

J. V. Shinde

Jayashri V Shinde completed BE and ME in computer engineering from the University of Pune and SRTM University, Nanded, India, respectively and pursuing PhD She is currently working as an assistant professor at LGNS College of Engineering at Nasik, India. Her research interests cover signal and image processing and computer networks.

Y. V. Joshi

Y V Joshi has completed BE (1986) and ME (1991) both from SGGS IE&T, Nanded and received his PhD from Indian Institute of Technology, Delhi, India in 1998. Currently, he is working as a director of Institute & professor of Electronics and Telecommunication Engineering at SGGS IE & T. He has published 20 international journal papers and presented at over 50 national and international conferences. He visited USA (2016), Dubai (2015), Malaysia and Singapore (2013), and China and HongKong (2016) for presenting papers in international conferences. He is a member of IEEE, Fellow of Institution of Engineers (India), Fellow of IETE (India), and a life member of ISTE. Email: [email protected]

R. R. Manthalkar

R R Manthalkar has completed BE (1988) and ME (1994) in electronics engineering from Marathwada University, Aurangabad, India. He received his PhD from Indian Institute of Technology, Kharagpur, India in 2003 with specialization in invariant texture classification. Currently, he is with the Department of Electronics and Telecommunication Engineering, SGGSIE & T, Nanded as a professor. His research interests cover signal and image processing, pattern recognition, digital VLSI & computer networks with over 25 international journal papers and more than 50 national and international conference paper publications. He is a life member of ISTE, Biomedical Engineering Society of India, System Society of India and Indian Society for continuing Engineering Education. Email: [email protected]