Content and Colour Distillation for Learning Image Translations with the Spatial Profile Loss
This addresses the need for simpler and more efficient training in image translation tasks, potentially reducing computational overhead and complexity for researchers and practitioners in computer vision.
The paper tackles the problem of image translation by proposing a novel loss function that directly computes shape/content and color/style differences between source and target images, enabling successful training without additional networks like discriminators. The method demonstrates effectiveness in tasks such as domain mapping, super-resolution, and makeup transfer, synthesizing realistic images as shown in extensive evaluations.
Generative adversarial networks has emerged as a defacto standard for image translation problems. To successfully drive such models, one has to rely on additional networks e.g., discriminators and/or perceptual networks. Training these networks with pixel based losses alone are generally not sufficient to learn the target distribution. In this paper, we propose a novel method of computing the loss directly between the source and target images that enable proper distillation of shape/content and colour/style. We show that this is useful in typical image-to-image translations allowing us to successfully drive the generator without relying on additional networks. We demonstrate this on many difficult image translation problems such as image-to-image domain mapping, single image super-resolution and photo realistic makeup transfer. Our extensive evaluation shows the effectiveness of the proposed formulation and its ability to synthesize realistic images. [Code release: https://github.com/ssarfraz/SPL]