Mismatch between timestamp.txt and number of images in HSERGB dataset?

[agirbau]

Hello @danielgehrig18!

While I was trying to evaluate your code ( #35 ) I realized that the timestamps.txt for some sequences (see baloon_popping) contain less timestamps (284) that the amount of frames in the images_corrected folder (334). Can you confirm if this is the correct behavior?

Thank you!

Edit: I'm seeing now that it's due to the number of event files in the events_aligned folder. Shouldn't these be equal to the amount of frames - 1 anyway?

[agirbau]

Ok, my bad. Previously I didn't understand the sentence in README: Note that in some folders there are more image files than event files. However, the image stamps in timestamp.txt should match with the event files and the additional images can be ignored. Now I do.