Save signal in WFDB Format 16

70 views
Skip to first unread message

Cuong Nguyen

unread,
Apr 1, 2024, 12:22:59 PMApr 1
to physionet-challenges
Hi Matt,

I have a question regarding the digitization challenge. I tried extracting the paper waveform, digitizing it using the grid voltage scale on paper (10mm for 1 mV). The digitized signal consisted of floating point values, mostly between 0.0 and 1.0.

In the `helper_code.py` script, the signal is cased to 16-bit integer (signal = np.asarray(signal, dtype=np.int16)) before saving to WFDB format. That rounded my digitized signal to have mostly zeroes. 

I wonder what I should do in this case. Any suggestions/related docs are appreciated.

Best, 
Cuong Nguyen

PhysioNet Challenge

unread,
Apr 1, 2024, 12:24:21 PMApr 1
to physionet-challenges
Dear Cuong,

For the unofficial phase, we are asking teams to "fill-in" the Waveform Database (WFDB) files that store the ECG time series:
https://physionet.org/physiotools/wag/header-5.htm

WFDB files and many other formats record ECGs signals as fixed-point numbers. They use integers for the time series with baselines/offsets and gains to shift and scale them, respectively, from digital to physical representations.

For example, if the header says "1000.0(0)/mV", then there is an analog-to-digital (ADC) gain of 1000 digital units per millivolt. In this case, you would multiple your signal values by 1000 before saving them. As you noticed, you can check that you created the data correctly by trying to load it again.

For the official phase, we will make this part of the Challenge more accessible, so feedback continues to be welcome.

Best,
Matt
(On behalf of the Challenge team.)

Please post questions and comments in the forum. However, if your question reveals information about your entry, then please email info at physionetchallenge.org. We may post parts of our reply publicly if we feel that all Challengers should benefit from it. We will not answer emails about the Challenge to any other address. This email is maintained by a group. Please do not email us individually.
Reply all
Reply to author
Forward
0 new messages