Improved Electrode Motion Artefact Denoising in ECG Using Convolutional Neural Networks and a Custom Loss Function.
Eoin BrophyBryan M. HennellyMaarten De VosGeraldine B. BoylanTomás WardPublished in: IEEE Access (2022)
Keyphrases
- loss function
- denoising
- convolutional neural networks
- pairwise
- logistic regression
- support vector
- image denoising
- image sequences
- risk minimization
- learning to rank
- image processing
- motion estimation
- convolutional network
- boosting framework
- empirical risk
- optical flow
- reproducing kernel hilbert space
- natural images
- hinge loss
- total variation
- regularization term
- gaussian process
- boosting algorithms
- multiresolution
- solution path
- special case
- feature selection
- convex loss functions