The scalographic pattern of Morlet continuous wavelet transform can differentiate bileaflet valve function

The authors have discussed the significance of the Morlet continuous wavelet transform of bileaflet mechanical heart valve (BLMHV) sound for detecting its malfunction: consecutive single patterns on the scalogram alway suggested its malfunction, whereas the tandem pattern with two steepled figures was demonstrated in both normal and malfunctioning valves. Therefore, authors have tried to distinguish this pattern between them by manually calculated multiple scalographic parameters. Although only the sum of wavelet coefficients (SWC) is supposed to be closer to valve sound property than other parameters, its calculation was not available in the original wavelet application. Therefore, the application was customized in the current study to semi-automatically calculate the SWC ratio between two figures for classifying the scalographic pattern of malfunctioning valves, and its efficacy to distinguish valve function was compared to other parameters. Among 155 BLMHVs, 6 valves with consecutive single patterns (type-I) and other 6 with two similar needle-like narrow figures (type-II) were confirmed to be a malfunction by fluoroscopic examination, whereas 14 malfunctioning valves with the tandem pattern which showed a great difference between two figure sizes (type-III) were distinguished from 129 normal valves by the cutoff point of the SWC ratio < 0.482 with the highest AUC (0.960) compared to other parameters by the ROC analysis.