Image-to-Image Translation with Conditional Adversarial Networks
Преобразование изображений в изображения с помощью условных состязательных сетей
2017-07-01
SCID: 54.1/9k34yybx
Discuss with AI
conditional adversarial networksgenerative adversarial networksimage synthesisimage-to-image translationsemantic label maps
Figures from the paper
Abstract (AI)
We investigate conditional adversarial networks as a general-purpose solution to image-to-image translation problems. These networks not only learn the mapping from input image to output image, but also learn a loss function to train this mapping. This makes it possible to apply the same generic approach to problems that traditionally would require very different loss formulations. We demonstrate that this approach is effective at synthesizing photos from label maps, reconstructing objects from edge maps, and colorizing images, among other tasks. Moreover, since the release of the pi×2pi× software associated with this paper, hundreds of twitter users have posted their own artistic experiments using our system. As a community, we no longer hand-engineer our mapping functions, and this work suggests we can achieve reasonable results without handengineering our loss functions either.
Key Findings
1
A single generic approach can handle tasks traditionally requiring different loss formulations, reducing the need for hand-engineered objectives.
2
Conditional adversarial networks provide a general-purpose framework for image-to-image translation by learning both the input-output mapping and its training loss.
3
The associated software enabled hundreds of Twitter users to create artistic experiments, indicating broad practical and community engagement.
4
The method effectively synthesizes photos from label maps, reconstructs objects from edge maps, and colorizes images.
5
The work suggests reasonable image-translation results can be achieved without manually engineering mapping functions or loss functions.
Research Object
conditional adversarial networks for image-to-image translation
Research Subject
generic image-to-image mapping and learned loss-function effectiveness across image synthesis, reconstruction, and colorization tasks
Publication Details
Publication Date
2017-07-01
Journal
Publisher
ISSN
Cited by
24258
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest
References available in scid.ai6
Adam: A Method for Stochastic Optimization2014
Reducing the Dimensionality of Data with Neural Networks2006
Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network2017
Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks2015
Image Style Transfer Using Convolutional Neural Networks2016
Data Driven VRP: A Neural Network Model to Learn Hidden Preferences for VRP2021
Cited by16
Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks2017
Least Squares Generative Adversarial Networks2017
Generative Adversarial Networks: An Overview2018
Scalable Diffusion Models with Transformers2023
Self-Attention Generative Adversarial Networks2018
Recent advances and applications of machine learning in solid-state materials science2019
Uformer: A General U-Shaped Transformer for Image Restoration2022
Vision Transformers for Single Image Dehazing2023
Artificial intelligence in the creative industries: a review2021
Deep Learning for Anomaly Detection in Time-Series Data: Review, Analysis, and Guidelines2021
AI-generated characters for supporting personalized learning and well-being2021
Artificial intelligence and machine learning for medical imaging: A technology review2021
A survey on data‐efficient algorithms in big data era2021
Transfer Learning from Synthetic to Real LiDAR Point Cloud for Semantic Segmentation2022
Histopathology in focus: a review on explainable multi-modal approaches for breast cancer diagnosis2024
High-resolution reconstruction of 3D seismic data using generative adversarial networks with global grouped coordinate attention mechanism2026