Image-to-Image Translation with Conditional Adversarial Networks

• BAIR, UC Berkeley ### Background

Video-to-Video Synthesis

• Nvidia, MIT, 2018

Background

• Image-to-Image Translation
  • Without modeling temporal dynamics, directly applying existing image synthesis approaches to an input video often results in temporally incoherent videos of low visual quality.

Main Strategy

• with a spatio-temporal adversarial objective
  • Through carefully-designed generator and discriminator architectures, coupled with a spatio-temporal adversarial objective, we achieve high-resolution, photorealistic, temporally coher- ent video results on a diverse set of input formats including segmentation masks, sketches, and poses.
• future video prediction
  • Finally, we apply our approach to future video prediction, outperforming several state-of-the-art competing systems.