Fetching the paper…

DDLP: Unsupervised Object-Centric Video Prediction with Deep Dynamic Latent Particles · Around