Sparsity-constrained one-dimensional convolutional autoencoder for reliable photoplethysmography template quality...

Photoplethysmography (PPG) signal quality assessment is essential for ensuring reliable physiological analysis, particularly in real-world conditions where motion artifacts and pathological variations are common. This study proposes a lightweight one-dimensional convolutional autoencoder for automatic classification of PPG signal quality based on template-level reconstruction error. The method incorporates template-based segmentation and spline interpolation for baseline correction, enabling consistent signal representation prior to model inference. The proposed approach is evaluated on a dataset containing both healthy and arrhythmic signals, providing a more realistic assessment compared to conventional studies. Experimental results demonstrate strong performance, achieving an accuracy of 96%, precision of 98%, recall of 98%, andF1-score of 98%. The model shows a high capability to preserve valid PPG signals while effectively rejecting corrupted segments. In addition, the proposed method shows improved performance compared with the skewness-based signal quality index, particularly for arrhythmic PPG signals. The model maintains a compact architecture with approximately 153 729 parameters (0.6 MB), making it suitable for deployment on resource-constrained edge devices. Analysis of the latent space reveals stable and structured feature representations, although partial overlap between healthy and arrhythmic samples is observed due to the physiological nature of rhythm irregularities. Overall, the proposed method provides an efficient solution for PPG signal quality assessment.