GANHopper: Multi-hop GAN for Unsupervised Image-to-Image Translation

Lira, Wallace; Merz, Johannes; Ritchie, Daniel; Cohen–Or, Daniel; Zhang, Hao

doi:10.1007/978-3-030-58574-7_22

Cited by 30 publications

(10 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Neural approach (joint generation w/ refinement): On the iterative sample refinement, GANHopper is the closest to ours in spirit [19]. The key difference is that our system produces a structured model (as opposed to a raster image) and the training process is non-sequential.…”

Section: Related Workmentioning

confidence: 99%

House-GAN++: Generative Adversarial Layout Refinement Networks

Nauata,

Hosseini,

Chang

et al. 2021

Preprint

View full text Add to dashboard Cite

Figure 1. The paper makes a breakthrough in the task of automated house layout generation. The right shows the mix of a ground-truth design made by an architect and our generated samples, based on the input bubble-diagram. Can you tell which one is the ground-truth? See the end of the caption for the answer. The paper proposes a novel generative adversarial layout refinement network, whose generator is trained to repeatedly apply and refine the design towards perfection. (The second sample from the right is the ground-truth.

show abstract

Section: Related Workmentioning

confidence: 99%

House-GAN++: Generative Adversarial Layout Refinement Networks

Nauata,

Hosseini,

Chang

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…The main idea is that a source image is first translated into the target domain and then translated back to the source domain, and the distance between the source and reconstructed images should be consistent at pixel-level. It becomes a popular strategy to solve the unpaired image translation problem [4,5,6,7,8,9]. However, cycle consistency maintains image content globally, causing poor performance on contentrich images.…”

Section: Image Translation Based On Cycle Consistencymentioning

confidence: 99%

“…Recent approaches [4,5,6,7,8,9,10] exploit cycle consistency [11] to loose the constraint of paired data for training, thus achieving better results. They assume that the translated image could be translated back to the source domain, and the reconstructed image should retain the same content with the original image.…”

Section: Introductionmentioning

confidence: 99%

Label-Free Regional Consistency for Image-to-Image Translation

Guo

Zhou

Zhou³

et al. 2021

2021 IEEE International Conference on Multimedia and Expo (ICME)

View full text Add to dashboard Cite

Image-to-Image translation aims to translate images from one domain to another. Existing approaches mainly stylize the images globally, while the local consistency between regions has been under-explored. Some instance-aware methods capture the regional consistency but heavily depend on wellannotated labels of a large-scale dataset. Besides, we observe that content-alike regions should have similar style between the target and translated images, however, little attention has been paid to explore such intrinsic property as explicit prior knowledge to guide the image translation process. In this paper, we aim to explore the label-free regional consistency for image-to-image translation. We propose regional relation consistency not only to maintain the global structure but also to keep a close look at the regional consistency, thus achieving more rigorous preservation of image contents. Moreover, we employ the phase of images as a semantic prior to select regions with similar content. We present phase-guided amplitude consistency to perform a more efficient local stylization. Extensive experiments verify that our approach outperforms the existing methods with a clear margin.

show abstract

“…Alternatively, spatial attention was exploited to drive better the adversarial training on unrealistic regions [43]. Some methods focus instead on generating intermediate representations of source and target [22,44] or continuous translations [61,48]. In the recent [20], authors exploit similarity with retrieved images to increase translation quality.…”

Section: Image-to-image Translationmentioning

confidence: 99%

Physics-informed Guided Disentanglement in Generative Networks

Pizzati¹,

Cerri²,

Charette³

2021

Preprint

View full text Add to dashboard Cite

Image-to-image translation (i2i) networks suffer from entanglement effects in presence of physics-related phenomena in target domain (such as occlusions, fog, etc), thus lowering the translation quality and variability. In this paper, we present a comprehensive method for disentangling physics-based traits in the translation, guiding the learning process with neural or physical models. For the latter, we integrate adversarial estimation and genetic algorithms to correctly achieve disentanglement. The results show our approach dramatically increase performances in many challenging scenarios for image translation. Keywords Adversarial learning • image to image translation • feature disentanglement • adverse weather • vision and rain • GAN

show abstract

GANHopper: Multi-hop GAN for Unsupervised Image-to-Image Translation

Cited by 30 publications

References 15 publications

House-GAN++: Generative Adversarial Layout Refinement Networks

House-GAN++: Generative Adversarial Layout Refinement Networks

Label-Free Regional Consistency for Image-to-Image Translation

Physics-informed Guided Disentanglement in Generative Networks

Contact Info

Product

Resources

About