How to use RNN with images

Hello, I found this paper here:
It seems like in order to deal with different ecg lengths, it deals with it by splitting each to 5s segments.
I have two questions regarding this:

During training, does each batch contain the segments of a single ecg?
What is the input (following the documentation notation (L,N,Hin​)) in the RNN in such cases?

For question 2, here is how I interpret it.
The output of the FC layer is 1024 thus the input size is 1024(corresponding to 5s of ecg), batch size = 1 if we are using a single ecg per batch and the sequence length is the number of 5s segments per ecg.

Thank you in advance.