Unsupervised Latent Space Translation Network
This work addresses domain adaptation in computer vision, offering an incremental improvement over existing methods for researchers in this field.
The paper tackled the problem of image-to-image translation by enhancing the UNIT framework to address its main drawbacks, specifically by introducing an additional adversarial discriminator on the latent representation to enforce similarity between domain distributions, resulting in greatly outperforming competing approaches on MNIST and USPS domain adaptation tasks.
One task that is often discussed in a computer vision is the mapping of an image from one domain to a corresponding image in another domain known as image-to-image translation. Currently there are several approaches solving this task. In this paper, we present an enhancement of the UNIT framework that aids in removing its main drawbacks. More specifically, we introduce an additional adversarial discriminator on the latent representation used instead of VAE, which enforces the latent space distributions of both domains to be similar. On MNIST and USPS domain adaptation tasks, this approach greatly outperforms competing approaches.