Expected image/EKG formats
98 views
Skip to first unread message
Sam J
Jun 18, 2024, 4:47:45 PMJun 18
to physionet-challenges
Hello Matt,
Thank you to all the challenge team for all the tremendous work put in. Ee really appreciate it.
I have a few questions regarding the expected format of the EKGs in the validation/hidden sets. I apologize if any of these questions were already clearly answered previously.
1- Can we assume that the EKG signal will be displayed for exactly the standard 10 seconds as expected on a standard 12-lead EKG?
2- Can we assume the standard 3x4 grid layout of the shorter 2.5 sec 12 EKG leads?
3- Can we assume that there will only be 1 rhythm lead (ie the longer lead displayed for the full 10 seconds)? or can there be a arbitrary number of rhythm leads?
4- Although is it most common for the rhythm lead to be "lead II", I assume that any other lead may be chosen in the validation set. Is that correct?
5- My understanding is that many different DPI values may be chosen to generate the validation images. Is it possible to have an idea of the range of DPI we should expect in the validation/hidden set?
Thank you again,
ECGénie team.
PhysioNet Challenge
Jun 18, 2024, 4:49:36 PMJun 18
to physionet-challenges
Hi Sam,
Thanks for the kind words and for your team's continued engagement with the Challenge.
The general answer to your question is that, yes, you can expect standard ECG papers, but standards also vary, and it's best to extract information from the ECG images whenever possible. For example, the course grid lines for the ECG papers are 5mm x 5mm and represent 0.2 seconds in time and 0.5 mV in amplitude, which should help with the signal duration, for example.
In particular:
1. We will provide the duration, i.e., the sampling frequency and number of samples, for each signal in the incomplete WFDB header files (scroll down to see what they look like):
https://moody-challenge.physionet.org/2024/#data-formats
Of course, ECG signals can be longer than 10 seconds. We will plot each full 10-second ECG segment from a record on a separate image. The time is given horizontally, and the different short and long ECG segments vary vertically.
2. Yes, you can assume the standard 3x4 grid layout of short ECG segments.
3. There may be different numbers of rhythm or long ECG segments; the Challenge website has a picture of a real-world example with multiple rhythm leads.
4. Yes, the rhythm lead or leads may be different leads, but lead II is a common choice.
5. The DPI/PPI values may vary, but you can expect that they will be large enough so that you're able to try to digitize or classify the ECG images; we're not trying to assess your ability to impute large amounts of missing data from low-resolution images.
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.
Thanks for the kind words and for your team's continued engagement with the Challenge.
The general answer to your question is that, yes, you can expect standard ECG papers, but standards also vary, and it's best to extract information from the ECG images whenever possible. For example, the course grid lines for the ECG papers are 5mm x 5mm and represent 0.2 seconds in time and 0.5 mV in amplitude, which should help with the signal duration, for example.
In particular:
1. We will provide the duration, i.e., the sampling frequency and number of samples, for each signal in the incomplete WFDB header files (scroll down to see what they look like):
https://moody-challenge.physionet.org/2024/#data-formats
Of course, ECG signals can be longer than 10 seconds. We will plot each full 10-second ECG segment from a record on a separate image. The time is given horizontally, and the different short and long ECG segments vary vertically.
2. Yes, you can assume the standard 3x4 grid layout of short ECG segments.
3. There may be different numbers of rhythm or long ECG segments; the Challenge website has a picture of a real-world example with multiple rhythm leads.
4. Yes, the rhythm lead or leads may be different leads, but lead II is a common choice.
5. The DPI/PPI values may vary, but you can expect that they will be large enough so that you're able to try to digitize or classify the ECG images; we're not trying to assess your ability to impute large amounts of missing data from low-resolution images.
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