the question about max sample.

[liqier]

From your paper we can see 'The max number of frames in each sample is 300. For samples with less than 300 frames, we repeat the samples until it reaches 300 frames.'
How do your method deal with the spatial and temporal discontinuities at the end of a sequence and the start of the same sequence as those would essentially be adjacent frames in this padding scheme? What happens if the authors simply use zero-padding instead?