Knee Joint Vibration Signal Analysis with Matching Pursuit Decomposition and Dynamic Weighted Classifier Fusion

Abstract

Analysis of knee joint vibration (VAG) signals can provide quantitative indices for detection of knee joint pathology at an early stage. In addition to the statistical features developed in the related previous studies, we extracted two separable features, that is, the number of atoms derived from the wavelet matching pursuit decomposition and the number of significant signal turns detected with the fixed threshold in the time domain. To perform a better classification over the data set of 89 VAG signals, we applied a novel classifier fusion system based on the dynamic weighted fusion (DWF) method to ameliorate the classification performance. For comparison, a single leastsquares support vector machine (LS-SVM) and the Bagging ensemble were used for the classification task as well. The results in terms of overall accuracy in percentage and area under the receiver operating characteristic curve obtained with the DWF-based classifier fusion method reached 88.76% and 0.9515, respectively, which demonstrated the effectiveness and superiority of the DWF method with two distinct features for the VAG signal analysis.

Feature Computing and Signal Classifications

“…The classification results of the Bayesian decision rule based on these two fluctuation features (accuracy: 88 %, A z : 0.957) [46] were better than the same classifier with the input features of FF and VMS (accuracy: 86.67 %, A z : 0.9096) [45]. To improve the classification performance, Cai et al [2] proposed the dynamic weighted classifier fusion system to distinguish VAG signals. The dynamic weighted classifier fusion system only with the number of wavelet matching pursuit decomposition and turns count with the fixed threshold as the input features was able to provide an overall accuracy of 88.76 % and the A z value of 0.9515 under the ROC curve [2].…”

“…To improve the classification performance, Cai et al [2] proposed the dynamic weighted classifier fusion system to distinguish VAG signals. The dynamic weighted classifier fusion system only with the number of wavelet matching pursuit decomposition and turns count with the fixed threshold as the input features was able to provide an overall accuracy of 88.76 % and the A z value of 0.9515 under the ROC curve [2]. A combination of most informative features can also help a classifier improve the diagnostic performance.…”

“…In order to utilize the diverse knowledge generated by the CSVMs, we use a dynamic weighted fusion (DWF) rule that can adaptively combine the CSVMs in the classifier fusion system [2,40].…”

“…The local decision generated by the kth CSVM is denoted as g k .f n /, with regards to the feature vector of the nth VAG signal, f n . The classifier fusion system then provides the overall classification decision, g DW F .f n /, by linearly combining the CSVMs with the weights w k .f n / that are varied from one signal to another [2]. Thus the DWF-based ensemble output can be formulated as [2] …”

Classification of Knee Joint Vibration Signals Using Bivariate Feature Distribution Estimation and Maximal Posterior Probability Decision Criterion

“…With the statistical features extracted [12][13][14][15], computational algorithms can be utilized to distinguish anomalous patterns associated with pathology [10,[16][17][18]. The aim of the present study is to use the kernel density estimation method with two-dimensional Gaussian kernels to represent the knee joint VAG signals in the bivariate feature space.…”