Reusing Discriminators for Encoding: Towards Unsupervised Image-to-Image Translation

@article{Chen2020ReusingDF,
  title={Reusing Discriminators for Encoding: Towards Unsupervised Image-to-Image Translation},
  author={Runfa Chen and Wen-bing Huang and Bing-hui Huang and Fuchun Sun and Bin Fang},
  journal={2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2020},
  pages={8165-8174}
}
  • Runfa Chen, W. Huang, +2 authors Bin Fang
  • Published 29 February 2020
  • Computer Science
  • 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
Unsupervised image-to-image translation is a central task in computer vision. Current translation frameworks will abandon the discriminator once the training process is completed. This paper contends a novel role of the discriminator by reusing it for encoding the images of the target domain. The proposed architecture, termed as NICE-GAN, exhibits two advantageous patterns over previous approaches: First, it is more compact since no independent encoding component is required; Second, this plug… 
Independent Encoder for Deep Hierarchical Unsupervised Image-to-Image Translation
TLDR
A novel architecture, termed as IEGAN, is proposed, which removes the encoder of each network and introduces an encoder that is independent of other networks, which can guide the model to translate high-quality images under the task with significant shape or texture change.
Unsupervised Generative Adversarial Networks with Cross-model Weight Transfer Mechanism for Image-to-image Translation
  • Xuguang Lai, Xiuxiu Bai, Yongqiang Hao
  • Computer Science
    2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW)
  • 2021
TLDR
A new model CWT-GAN is proposed, which introduces the cross-model weight transfer mechanism, and performs better than the state-of-the-art methods in a single translation direction for several datasets.
DeepI2I: Enabling Deep Hierarchical Image-to-Image Translation by Transferring from GANs
TLDR
A novel deep hierarchical Image-to-Image Translation method, called DeepI2I, that transfers knowledge from pre-trained GANs and qualitatively and quantitatively demonstrates that transfer learning significantly improves the performance of I2I systems, especially for small datasets.
Global and Local Alignment Networks for Unpaired Image-to-Image Translation
TLDR
A novel approach, Global and Local Alignment Networks (GLA-Net), which aims to transfer the input image from the source domain to the target domain and introduces a local alignment network, which employs a pretrained self-supervised model to produce an attention map via a novel local alignment loss.
Unaligned Image-to-Image Translation by Learning to Reweight
TLDR
This paper proposes to select images based on importance reweighting and develop a method to learn the weights and perform translation simultaneously and automatically and compares the proposed method with state-of-the-art image translation approaches and presents qualitative and quantitative results on different tasks with unaligned domains.
Powers of layers for image-to-image translation
TLDR
A simple architecture to address unpaired image-to-image translation tasks: style or class transfer, denoising, deblurring, deblocking, etc, which is comparable or better than CycleGAN with significantly fewer parameters.
TransferI2I: Transfer Learning for Image-to-Image Translation from Small Datasets
TLDR
This paper decouple the learning process into the image generation step and the I2I translation step, and proposes two novel techniques: source-target initialization and self-initialization of the adaptor layer.
Complementary, Heterogeneous and Adversarial Networks for Image-to-Image Translation
TLDR
It is argued that heterogeneous generators are complementary to each other and will benefit the generation of images, and a novel approach to integrate multi-level and multi-scale features in the discriminator is proposed.
DINO: A Conditional Energy-Based GAN for Domain Translation
TLDR
This work proposes an alternative method for conditioning and presents a new framework, where two networks are simultaneously trained, in a supervised manner, to perform domain translation in opposite directions, which is not only better at capturing the shared information between two domains but is more generic and can be applied to a broader range of problems.
GAN generation of synthetic multispectral satellite images
TLDR
The unpaired style transfer method is used to modify images that are dominant in vegetation land cover into image that are dominated by bare land cover and vice versa, showing that the land-cover transfer is an easier task compared to the GAN generation from scratch.
...
1
2
3
...

References

SHOWING 1-10 OF 44 REFERENCES
TraVeLGAN: Image-To-Image Translation by Transformation Vector Learning
TLDR
This work introduces a novel GAN based on preserving intra-domain vector transformations in a latent space learned by a siamese network that helps generators learn mappings between more complex domains that differ from each other by more than just style or texture.
DualGAN: Unsupervised Dual Learning for Image-to-Image Translation
TLDR
A novel dual-GAN mechanism is developed, which enables image translators to be trained from two sets of unlabeled images from two domains, and can even achieve comparable or slightly better results than conditional GAN trained on fully labeled data.
XGAN: Unsupervised Image-to-Image Translation for many-to-many Mappings
TLDR
XGAN ("Cross-GAN"), a dual adversarial autoencoder, is introduced, which captures a shared representation of the common domain semantic content in an unsupervised way, while jointly learning the domain-to-domain image translations in both directions.
ComboGAN: Unrestrained Scalability for Image Domain Translation
TLDR
This paper proposes a multi-component image translation model and training scheme which scales linearly - both in resource consumption and time required - with the number of domains and demonstrates its capabilities on a dataset of paintings by 14 different artists.
Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks
TLDR
This work presents an approach for learning to translate an image from a source domain X to a target domain Y in the absence of paired examples, and introduces a cycle consistency loss to push F(G(X)) ≈ X (and vice versa).
Multimodal Unsupervised Image-to-Image Translation
TLDR
A Multimodal Unsupervised Image-to-image Translation (MUNIT) framework that assumes that the image representation can be decomposed into a content code that is domain-invariant, and a style code that captures domain-specific properties.
TransGaGa: Geometry-Aware Unsupervised Image-To-Image Translation
TLDR
A novel disentangle-and-translate framework to tackle the complex objects image-to-image translation task, which disentangles image space into a Cartesian product of the appearance and the geometry latent spaces and supports multimodal translation.
Unsupervised Image-to-Image Translation Networks
TLDR
This work makes a shared-latent space assumption and proposes an unsupervised image-to-image translation framework based on Coupled GANs that achieves state-of-the-art performance on benchmark datasets.
U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation
TLDR
A novel method for unsupervised image- to-image translation, which incorporates a new attention module and a new learnable normalization function in an end-to-end manner, which can translate both images requiring holistic changes and images requiring large shape changes.
StarGAN: Unified Generative Adversarial Networks for Multi-domain Image-to-Image Translation
TLDR
A unified model architecture of StarGAN allows simultaneous training of multiple datasets with different domains within a single network, which leads to StarGAN's superior quality of translated images compared to existing models as well as the novel capability of flexibly translating an input image to any desired target domain.
...
1
2
3
4
5
...