Image-to-Image Translation with Conditional Adversarial Networks
This work provides a general-purpose solution for image-to-image translation, benefiting researchers and practitioners in computer vision by automating loss function design, though it builds incrementally on existing adversarial network methods.
The paper tackles image-to-image translation problems by using conditional adversarial networks, which learn both the mapping from input to output images and a loss function, eliminating the need for hand-engineered loss formulations. It demonstrates effectiveness in tasks like synthesizing photos from label maps, reconstructing objects from edge maps, and colorizing images, with widespread adoption by users such as artists showing its applicability.
We investigate conditional adversarial networks as a general-purpose solution to image-to-image translation problems. These networks not only learn the mapping from input image to output image, but also learn a loss function to train this mapping. This makes it possible to apply the same generic approach to problems that traditionally would require very different loss formulations. We demonstrate that this approach is effective at synthesizing photos from label maps, reconstructing objects from edge maps, and colorizing images, among other tasks. Indeed, since the release of the pix2pix software associated with this paper, a large number of internet users (many of them artists) have posted their own experiments with our system, further demonstrating its wide applicability and ease of adoption without the need for parameter tweaking. As a community, we no longer hand-engineer our mapping functions, and this work suggests we can achieve reasonable results without hand-engineering our loss functions either.