| Abstrakt: |
In order to detect variations in blood volume in the peripheral arterial pulse, photoplethysmography (PPG) can be used as an electro-optical method. Recognized as a straightforward, non-invasive, and reasonably priced method for diagnosing cardiovascular issues, PPG pulse characterization has attracted a lot of attention in recent years. IoT based analysis of PPG contours can shed light on cardiac characteristics at various points in the cardiac cycle. Loss of pulsatility associated with age and Cardio Vascular Disease (CVD) is the primary limitation of contour analysis The e waves are classified through optimistic discrete wavelet and CNN based approach. As a result, IoT based accurate delineation of the PPG pulse is necessary for accurate detection of heart illness. In addition, resampling the SDPPG signal ensures the occurrence of particular facts when it comes to the process of building slabs of attention, in which the undesired slabs are primarily removed by means of onset criteria. Research provides a comparison of the performance of the proposed method to that of other machine learning based techniques. In terms of classification accuracy (Normal, P1, and P2 pulses: 95.9%, 93.4%, and 90.08% respectively), wavelet-based CNN with PPG signal outperforms CNN with PPG and ABP signals. Correct classification is aided by CNN's ability to extract many features associated with premature pulses. To ensure that no pulses are missed, our method estimates the wavelet transform at one-second intervals over the whole signal's duration. CNN's incremental fine tuning also aids in improving both sensitivity and specificity. The wavelet-based CNN outperforms other state-of-the-art approaches in terms of accuracy, sensitivity, and specificity when classifying waves. |
Full text from SpringerLink