A supervised machine learning semantic segmentation approach for detecting artifacts in plethysmography signals from wearables

Abstract

Objective. Wearable devices equipped with plethysmography (PPG) sensors provided a low-cost, long-term solution to early diagnosis and continuous screening of heart conditions. However PPG signals collected from such devices often suffer from corruption caused by artifacts. The objective of this study is to develop an effective supervised algorithm to locate the regions of artifacts within PPG signals. Approach. We treat artifact detection as a 1D segmentation problem. We solve it via a novel combination of an active-contour-based loss and an adapted U-Net architecture. The proposed algorithm was trained on the PPG DaLiA training set, and further evaluated on the PPG DaLiA testing set, WESAD dataset and TROIKA dataset. Main results. We evaluated with the DICE score, a well-established metric for segmentation accuracy evaluation in the field of computer vision. The proposed method outperforms baseline methods on all three datasets by a large margin (≈ 7 percentage points above the next best method). On the PPG DaLiA testing set, WESAD dataset and TROIKA dataset, the proposed method achieved 0.8734±0.0018, 0.9114±0.0033 and 0.8050±0.0116 respectively. The next best method only achieved 0.8068±0.0014, 0.8446±0.0013 and 0.7247±0.0050. Significance. The proposed method is able to pinpoint exact locations of artifacts with high precision; in the past, we had only a binary classification of whether a PPG signal has good or poor quality. This more nuanced information will be critical to further inform the design of algorithms to detect cardiac arrhythmia.

Photoplethysmogram Analysis and Applications: An Integrative Review

“…In their study, PPG quality was classified into five grades; it was found that the random forest SQI evaluation algorithm had the highest classification accuracy, with an accuracy of 74.5%. In a recent study, Guo et al (2021) detected wearable PPG artifacts with a DICE score of 0.87–0.91 through a combination of active-contour-based loss and an adapted U-Net architecture; compared to the existing general research methods, this method shows superior performance. However, to sufficiently verify the performance of the deep-learning model, verification using more abundant data is required.…”

Machine Learning Approaches of cotact PPG and its Potential Research Challanges

“…These algorithms can learn to identify patterns and relationships in the data that are difficult for humans to discern. For example, Machine Learning can detect subtle changes in breathing patterns that could indicate a respiratory disorder in respiratory plethysmography [9][10][11].…”

Wearable Photoplethysmography for Cardiovascular Monitoring

“…2(c) . Several approaches have been proposed, including statistical analysis of pulse wave shape [ 56 ], assessing the level of perfusion through the perfusion index (the ratio of the amplitude of the pulsatile component of the PPG to its baseline) [ 56 ], assessing the similarity of successive pulse wave shapes using template matching [ 86 ] or dynamic time warping [ 87 ], and deep learning [ 88 ]. For further details of techniques, see [ 10 ].…”