RISP: Rendering-Invariant State Predictor
with Differentiable Simulation and Rendering
for Cross-Domain Parameter Estimation
Pingchuan Ma*, Tao Du*, Joshua B. Tenenbaum, Wojciech Matusik, and Chuang Gan
International Conference on Learning Representations (ICLR), 2022 [Oral]
We use RISP to reconstruct the dynamic motion of a quadrotor (left) from a real-world video input (right).