Plantar fascia ultrasound images characterization and classification using support vector machine

Boussouar, A; Meziane, F; Walton, LA

Abstract

The examination of plantar fascia (PF) ultrasound (US) images is subjective and based on the visual perceptions and manual biometric measurements carried out by medical experts. US images feature extraction, characterization and classification have been widely introduced for improving the accuracy of medical assessment, reducing its subjective nature and the time required by medical experts for PF pathology diagnosis. In this paper, we develop an automated supervised classification approach using the Support Vector Machine (Linear and Kernel) to distinguishes between symptomatic and asymptomatic PF cases. Such an approach will facilitate the characterization and the classification of the PF area for the identification of patients with inferior heel pain at risk of plantar fasciitis. Six feature sets were extracted from the segmented PF region. Additionally, features normalization, features ranking and selection analysis using an unsupervised infinity selection method were introduced for the characterization and the classification of symptomatic and asymptomatic PF subjects.

The performance of the classifiers was assessed using confusion matrix attributes and some derived performance measures including recall, specificity, balanced accuracy, precision, F-score and Matthew’s correlation coefficient. Using the best selected features sets, Linear SVM and Kernel SVM achieved an F-Score of 97.06 and 98.05 respectively.

Citation

Boussouar, A., Meziane, F., & Walton, L. (2019, October). Plantar fascia ultrasound images characterization and classification using support vector machine. Presented at The 5th International Conference on Advanced Intelligent Systems and Applications, Cairo, Egypt

• Presentation Conference Type: Other
• Conference Name: The 5th International Conference on Advanced Intelligent Systems and Applications
• Conference Location: Cairo, Egypt
• Start Date: Oct 26, 2019
• End Date: Oct 28, 2019
• Acceptance Date: May 15, 2019
• Online Publication Date: Oct 2, 2019
• Publication Date: Oct 2, 2019
• Deposit Date: May 28, 2019
• Publicly Available Date: Oct 29, 2019
• Series Title: Advances in Intelligent Systems and Computing
• Series Number: 1058
• Book Title: Proceedings of the International Conference on Advanced Intelligent Systems and Informatics 2019
• ISBN: 9783030311285;-9783030311292
• DOI: https://doi.org/10.1007/978-3-030-31129-2_10
• Publisher URL: https://link.springer.com/chapter/10.1007/978-3-030-31129-2_10
• Related Public URLs: http://www.egyptscience.net/AISI2019/home.html
• Additional Information: Event Type : Conference