Questions about input data

[Kajtek Jot]

Dear Physionet Team,

While creating my algorithm I've read the Data Formatting section a couple times and after that two questions came to my mind. I am developing only the image classification task based on the images.

1) Is the size of input images in your hidden datasets constant? Or the resolution may vary depending on the origin of the ecg image?

2) On your website there are two places where it is written that "the classes are: normal and abnormal". However, ECG datasets usually provide more than normal/abnormal class (so it is in ptb-xl dataset). I understand that it doesn't matter whether I send a model that is trained on 2 or more output classes, but my question is: Does Physionet Team expect users to perform normal/abnormal classification or multi-class & multi-label classification with classes from ptb-xl dataset such as: norm, mi, sttc, cd, hyp?

Thanks in advance,

Kajetan.

[PhysioNet Challenge]

Dear Kajetan,

Thanks for your questions. We are working to launch the official phase in the coming days, but I am writing to provide quick answers to your questions.

The sizes, dimensions, resolution, and so on of the ECG images in the hidden validation and test sets may differ from those of the ECG images that we provided in the public training set. They may also have other differences such as different signal lengths and sampling rates and different image distortions. Your code will have access to ECG images with the default options for our ECG image generation code, but we encourage the teams to change the options to generate different ECG images for training their models.

We processed the PTB-XL database before running your code on it. These data processing steps update the format and classes of each database so that your code can expect the same format and classes from different databases. Please see this section of the README for details, and please run this script to prepare the data in the same format that we will use for the Challenge:
https://github.com/physionetchallenges/python-example-2024?tab=readme-ov-file#how-do-i-create-data-for-these-scripts

As always, we are planning to update the Challenge for the official phase to address the lessons learned during the unofficial phase, so some of these details may change shortly.

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.