Beat annotations in the Fantasia Database
kevinashaw opened this issue · comments
I am trying to read the Fantasia Database (link) including the beat annotations. For the record, I am able to read the ECG waveforms, just not the annotations.
The introduction for the dataset states, "Each heartbeat was annotated using an automated arrhythmia detection algorithm, and each beat annotation was verified by visual inspection." Which leads me to believe that beat annotations are available in the dataset.
When I inspect the files in the dataset, I find .hea, .dat and .ecg. From my research, none of these store beat annotations, with .hea and .dat being the MIT format and .ecg being the ISHNE Holter format. Furthermore, according to the documentation there would need to be a .atr for beat annotation.
Using the wfdb library with Python, I have tried to use rdann on both the .dat and the .ecg file and both fail. For example:
annotation = wfdb.rdann(filepathname, extension='ecg')
Is there a way to read the beat annotations from this data set?
Is it that these datasets do not include annotations, or am i incorrectly processing the files.
Thank you in advance.
-K
Further research showed that not only can rdann read .ecg files but that (contrary of the ISHNE Holter document) .ecg files can contain beat annotations.
For reasons that are not clear to me, the above listed python command does now work and I am now able to extract annotations from the .ecg files in the Fantasia dataset.
If I were to offer a recommendation, it would be helpful if the rdann documentation listed the file formats that it is capable of reading. It is always useful to know what types of file we should expect to be readable.
And finally, thank you for providing such amazing tools for working with biosignals!