Probably there isn't much difference! Captions are not always accurate, and the caption writing can be of a rather variable standard.
However, if I were a caption writer I would use DOG BARKS if there were a few woofs from ONE dog, and DOG BARKING if there were many woofs from ONE or MORE dogs. You see, Dog BARKING could mean there is ONE dog doing some continuous barking (the dog is barking), or it could be a compound noun meaning 'the barking sound that dogs make.'
Mind you, DOG BARKS could also be a compound noun. Dog barks (the sound). Dog bites (the injuries). A dog bite is painful. The dog bark is an instantly recognizable sound. So DOG BARKS is actually a little ambiguous too. Is it A dog barking. Or is it the sound of barking in general.
Sausage rolls. Does this mean a sausage is rolling? Or is it a sausage in pastry?
In summary, I think DOG BARKS is short for A dog IS barking now but not much! And DOG BARKING suggests a bit more barking, maybe from one dog, maybe from more than one.