Guest
This is a demo of our work on visual rhythm and beat: ...and also a look at what Nathan thinks about when he dances. (The explanation below makes references to our paper, which can be found at ) ---How it was made--- First I stabilized this video: Then I had to modify our dancification algorithm so that instead of traversing visual beats randomly, it follows the dance steps in “Gangnam Style.“ I started with the automatically detected visual beats, and adjusted them manually to follow a repeating pattern of 4 configurations (left foot up, both down, right foot up, both down ). Then I mapped the steps of gangnam style (right, left. right, right, left, right, etc...) to a pattern of these configurations and used that pattern as a constraint on the random walk described in the paper. Once that was done I could generate an infinite, non-repeating video of t
