Fig 1.
Adaptive noise filter concept.
Fig 2.
Block diagram of the proposed motion artefact reduction algorithm.
Fig 3.
Hampel filter based estimation.
Table 1.
List of the symbols used.
Fig 4.
Hampel filter-based estimation.
Fig 5.
Conceptual structure of hampel filter-based filtering.
Table 2.
MHEALTH datasets.
Fig 6.
The correlation between the acceleration signals and the ECG signal at different lags.
Fig 7.
The correlation between the ECG signal and the acceleration signal.
Fig 8.
Filtering results using the ECG signal recorded during walking.
Fig 9.
Filtering results using the ECG signal recorded during running.
Table 3.
Performance statistics of the adaptive filtering algorithms at different activity levels.
Fig 10.
Filtering results using the ECG signal recorded during free movement.
Fig 11.
Filtering results using the ECG signal recorded during knee bending exercise.
Fig 12.
Example of treatment of heavey contaminated ECG signal.
Fig 13.
Filtering results using the ECG signal recorded during walking following by jogging exercise.
Table 4.
Percentage of the removed motion artefact from the raw signal.
Table 5.
Performance statistics of the tested algorithms at different activity levels.
Table 6.
Statistical results of the filtering performance in terms of R-beak detection.
Fig 14.
Summary of the performance during different forms of exercise.
Fig 15.
The QRS complex detection accuracy before and after the filtering using the proposed methods.