Hi. I am a single track participant.
I have a question regarding the training phase.
I know that only medium exposed LDR image should be used in the inference phase,
Should I use only the medium exposed LDR image during training as in the inference phase?
Or, can we use short and long exposed LDR images together during training?
Thank you.
Posted by: Yein @ Feb. 22, 2021, 9:55 a.m.Hello Yein,
You raise an interesting point on using multi-image training for single-image inference. That could be on its own an interesting research question, however after careful consideration we decided not to allow multi-frame training for this first HDR single image challenge track, as the challenge description and the spirit of the track is to operate on a limited set-up where multi-image data is not be available at all, and making any change to that clear set-up could be confusing and unfair to some participants. We will further clarify this in the challenge description as soon as possible to fix any potential confusion on the current wording.
I would like to state again that despite not being allowed in this ongoing challenge, it is an interesting approach and is something to note down for the future.
Best regards,
Edu.