Loading the signal
The toolkit loads the ECG signal from the raw APEX recording. Use the file selector to find the correct .mat file.
Post-Processing
The toolkit offers a variety of post-processing options. Ensemble methods work directly on the ECG signal. On the other hand, all other methods function directly on the processed tachogram (inter-beat intervals).
Ensemble methods
The Ensemble processing unit compares each beat to the average beat to eliminate odd beats, based on correlation. The unit works as follows:
- For each detected R-peak, get a 200mn window of the raw ECG signal centered around the peak.
- Average all of these windows to get the shape of the average peak.
- Discard all peaks that have a correlation with the average beat lower than the specified threshold.
MAD Filtering
The Median Absolute Deviation Filter can be used to detect and remove outlier inter-beat intervals.
Missed Beat Detector
The missed beat detector passes over the signal to detect any potential missed beats. Missed beats are detected by looking for a sudden doubling of the RR-interval before a return to normal.
Tachogram
Interpolation
To generate the continuous heart rate signal over time, use one of the following methods:
- Direct (linear) interpolation
- Splines (cubic) interpolation, with a supplied smoothing coefficient.
Median Fitering
Once the final tachogram has been obtained, it can be smoothed using a median filter.
Results
Evaluation
The toolkit displays various quality metrics in the left pane after processing.
Graphs
Graphs are shown in the right-hand pane. They can be explored using the slider and zoom options.