CAFLOW: Conditional Autoregressive Flows
This work addresses image-to-image translation for computer vision applications, representing an incremental advancement in model design.
The paper tackles the problem of diverse image-to-image translation by introducing CAFLOW, a model that combines auto-regressive modeling with conditional normalizing flows, resulting in improved performance over previous conditional flow designs.
We introduce CAFLOW, a new diverse image-to-image translation model that simultaneously leverages the power of auto-regressive modeling and the modeling efficiency of conditional normalizing flows. We transform the conditioning image into a sequence of latent encodings using a multi-scale normalizing flow and repeat the process for the conditioned image. We model the conditional distribution of the latent encodings by modeling the auto-regressive distributions with an efficient multi-scale normalizing flow, where each conditioning factor affects image synthesis at its respective resolution scale. Our proposed framework performs well on a range of image-to-image translation tasks. It outperforms former designs of conditional flows because of its expressive auto-regressive structure.