Auto-Conditioned Recurrent Networks for Extended Complex Human Motion Synthesis
This addresses the challenge of generating realistic, long-duration human motions for applications in animation and simulation, representing a significant advance beyond previous methods limited to simple motions.
The paper tackles the problem of synthesizing extended, complex human motions by introducing an auto-conditioned Recurrent Neural Network (acRNN) that explicitly accommodates autoregressive noise accumulation, enabling the generation of over 18,000 continuous frames (300 seconds) of new motion for styles like dances and martial arts.
We present a real-time method for synthesizing highly complex human motions using a novel training regime we call the auto-conditioned Recurrent Neural Network (acRNN). Recently, researchers have attempted to synthesize new motion by using autoregressive techniques, but existing methods tend to freeze or diverge after a couple of seconds due to an accumulation of errors that are fed back into the network. Furthermore, such methods have only been shown to be reliable for relatively simple human motions, such as walking or running. In contrast, our approach can synthesize arbitrary motions with highly complex styles, including dances or martial arts in addition to locomotion. The acRNN is able to accomplish this by explicitly accommodating for autoregressive noise accumulation during training. Our work is the first to our knowledge that demonstrates the ability to generate over 18,000 continuous frames (300 seconds) of new complex human motion w.r.t. different styles.