Time-frequency based feature extraction for the analysis of vibroarthographic signals. (July 2018)

• Record Type: Journal Article
• Authors: Nalband, Saif; Valliappan, C.A.; Prince, A. Amalin; Agrawal, Anita
• Is Part Of: Computers & electrical engineering. Volume 69(2018)
• Page Start: 720
• Page End: 731
• Publication Date: 2018-07
• Subjects: Knee-joint disorders -- Vibroarthographic signals -- Time-frequency -- LS-SVM -- Biomedical signal processing; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2018.02.046
• Languages: English
• ISSNs: 0045-7906

Abstract

Highlights: We propose to develop a computer aided diagnosis system based on time-frequency analysis for diagnosing knee-joint disorders. Time-frequency analysis has been carried out using smoothed pseudo Wigner Ville distribution (SPWVD) and Hilbert–Huang transforms (HHT) with CEEMDAN for computing IMFs. The time-frequency representation is considered as a time-frequency image and statistical features are extracted. A pattern classification is performed to compare the effectiveness of proposed methods using LS-SVM. Abstract: In this study, we propose to develop a computer-aided diagnostic (CAD) system based on time-frequency analysis for the diagnosis of knee-joint disorders. Two methodologies based on nonstationary signal processing techniques have been proposed. We propose to use smoothed pseudo Wigner–Ville distribution (SPWVD) and a modified version of Hilbert–Huang transform (HHT) for the analysis of vibroarthographic (VAG) signals. Traditional HHT consists of empirical mode decomposition (EMD) for computing intrinsic mode functions (IMFs) and Hilbert transform (HT). But we propose to use complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) for computing IMFs. The time-frequency representation of the proposed methods is considered as a time-frequency image. Statistical features such as mean, standard deviation, skewness and kurtosis are extracted. A pattern classification is carried out using Least square support vector machine (LS-SVM) to compare performance. Results concluded that highest classification accuracy of 88.76% was obtained by features extracted from CEEMDAN-HHT.