TY  - JOUR
TI  - Image-to-Image Translation: pix2pix and CycleGAN
AU  - Isola, P.
AU  - Zhu, J.-Y.
AU  - Zhou, T.
AU  - Efros, A. A.
PY  - 2017
DA  - 2017///
JO  - CVPR 2017 / ICCV 2017 (IEEE/CVF)
AB  - Isola et al. (pix2pix, CVPR 2017) and Zhu et al. (CycleGAN, ICCV 2017) established the two canonical approaches to image-to-image translation. pix2pix trains a conditional GAN on aligned image pairs to learn a mapping between visual domains, such as edge maps to photographs or sketches to coloured artwork. CycleGAN removes the paired-data requirement by enforcing cycle consistency, learning to translate between domains from unaligned collections and enabling applications such as photo-to-anime and style-domain transfer. Together these papers underlie sketch colourisation, edge-to-image generation and concept-to-frame pipelines in animation production, and they are the paired-training ancestor that ControlNet explicitly cites. Both are foundational content in AI and animation courses.
KW  - generative-ai
KW  - image-generation
UR  - https://arxiv.org/abs/1611.07004
LA  - en
ER  -