Blood Cell Signal Classification via Hilbert-Huang Transform Combined with Information Entropy

Abstract

The new applications of Hilbert-Huang Transform (HHT) and information entropy are explored in the analysis and classification of Blood Cell Signal (BCS). The BCS is processed by the Empirical Mode Decomposition (EMD) of HHT, and then Intrinsic Mode Function (IMF) of BCS is obtained. The frequency feature of IMFs are analyzed by Fourier transform. Both the Hilbert marginal spectrums of BCS and IMFs of healthy people are compared with the patients’. The hilbert marginal spectrum entropy of 38 healthy people and 39 patients are extracted. Five sets of feature vectors in the classification experiments are stacked with the entropy of hilbert marginal spectrum described the nonlinear dynamics feature. The highest Classification Rate (CR) is 91.89%, meanwhile the good sensitivity and specificity are guaranteed. The method of Hilbert-Huang Transform Combined with Information Entropy is efficient and effective in extracting the BCS nonlinear dynamics features. Therefore, it is significative to assist BCS clinical analysis and processing.