Leveraging Local Domains for Image-to-Image Translation

Autor:	Dell'Eva, Anthony, Pizzati, Fabio, Bertozzi, Massimo, de Charette, Raoul
Rok vydání:	2021
Předmět:	Computer Science - Computer Vision and Pattern Recognition Computer Science - Artificial Intelligence Computer Science - Machine Learning Computer Science - Robotics
Druh dokumentu:	Working Paper
Popis:	Image-to-image (i2i) networks struggle to capture local changes because they do not affect the global scene structure. For example, translating from highway scenes to offroad, i2i networks easily focus on global color features but ignore obvious traits for humans like the absence of lane markings. In this paper, we leverage human knowledge about spatial domain characteristics which we refer to as 'local domains' and demonstrate its benefit for image-to-image translation. Relying on a simple geometrical guidance, we train a patch-based GAN on few source data and hallucinate a new unseen domain which subsequently eases transfer learning to target. We experiment on three tasks ranging from unstructured environments to adverse weather. Our comprehensive evaluation setting shows we are able to generate realistic translations, with minimal priors, and training only on a few images. Furthermore, when trained on our translations images we show that all tested proxy tasks are significantly improved, without ever seeing target domain at training. Comment: VISAPP 2022 Best Paper Award
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2109.04468 Zobrazit plný text záznamu View this record from Arxiv