A convolutional variational autoencoder trained only on clean ICU arterial blood pressure data detects waveform artefacts at about 90% sensitivity and specificity and outperforms PCA reconstruction.
Multifractal analysis of hemodynamic behavior: intraoperative instability and its pharmacological manipulation,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
DeepClean -- self-supervised artefact rejection for intensive care waveform data using deep generative learning
A convolutional variational autoencoder trained only on clean ICU arterial blood pressure data detects waveform artefacts at about 90% sensitivity and specificity and outperforms PCA reconstruction.