Smoothing the Disentangled Latent Style Space for Unsupervised Image-to-Image Translation
This work addresses a specific issue in image-to-image translation for computer vision applications, offering an incremental improvement by enhancing existing models with plug-in losses.
The paper tackles the problem of abrupt appearance changes and poor cross-domain interpolations in unsupervised image-to-image translation by proposing a training protocol with three losses to learn a smooth and disentangled latent style space, resulting in significantly improved image quality and interpolation graduality across different datasets.
Image-to-Image (I2I) multi-domain translation models are usually evaluated also using the quality of their semantic interpolation results. However, state-of-the-art models frequently show abrupt changes in the image appearance during interpolation, and usually perform poorly in interpolations across domains. In this paper, we propose a new training protocol based on three specific losses which help a translation network to learn a smooth and disentangled latent style space in which: 1) Both intra- and inter-domain interpolations correspond to gradual changes in the generated images and 2) The content of the source image is better preserved during the translation. Moreover, we propose a novel evaluation metric to properly measure the smoothness of latent style space of I2I translation models. The proposed method can be plugged into existing translation approaches, and our extensive experiments on different datasets show that it can significantly boost the quality of the generated images and the graduality of the interpolations.