For decades machines are ready to understand simple musical features like beats per minute. Now AI is boosting their abilities to the purpose that they will not only find out what particular genre of music is playing but also the way to appropriately dance
The researchers introduce a completely unique decomposition-to-composition framework which will transform basic body movements into complex dances conditioned on music. The decomposition phase learns the way to perform basic dance moves by explaining and normalizing dance units segments from a video of real dancing sequences by a kinematic beat detector. A music-to-movement Generative Adversarial Network (GAN) generates music on conditioned dance moves. Researchers then select style and beats, integrate dance units during a repetitive way and apply a “beat warper” to the created dance unit order to perform the final output dance.
Watch this video on youtube to understand how it works – Click Here!
AI researchers aren’t generally referred to as party animals, they are doing seem to possess a passion for dance. In October Synced published the story Shake Your Booty: AI Deepfakes Dance Moves From one Picture, which reports on ShanghaiTech University and Tencent AI Lab researchers’ 3D body mesh recovery module Liquid Warping GAN, which may deepfake dance moves from one picture. The study is presented within the paper Liquid Warping GAN: A Unified Framework for Human Motion Imitation, Appearance Transfer and Novel View Synthesis.
The August Synced story Silicon Night Fever: Berkeley AI Gets Down meanwhile introduces the UC Berkeley paper Everybody Dance Now, which proposes a video-to-video translation approach for dance moves.
(Image by C. Chan et. al., UC Berkeley)
As dependent on style and passion as it is a technique, the language of dance has never been clear cut for humans, so why should it be for machines?
Artificial intelligence is proving to be a very intelligent tool when it comes to handling videos of people. Face-swapping deepfakes have been the most obvious example, but new applications are being discovered every day. The latest? Call it deepfakes for dancing. It utilises AI to know someone’s dance moves and imitate them on to a target body.
This paper offers an easy method for “do as I do” motion transfer: given a source video of a person dancing, we can covert that performance to a novel (amateur) target after only some minutes of the target point making standard moves. To know more about watch this video out – Everybody Dance Now!
Tensorflow is an app that helps learned to dance using the AI.
Ever wondered of dancing at a weird place or forest? A photo can come live to places and dance with the app named Jiggy