Filters








745 Hits in 3.1 sec

Analyzing and Improving the Image Quality of StyleGAN [article]

Tero Karras, Samuli Laine, Miika Aittala, Janne Hellsten, Jaakko Lehtinen, Timo Aila
2020 arXiv   pre-print
Overall, our improved model redefines the state of the art in unconditional image modeling, both in terms of existing distribution quality metrics as well as perceived image quality.  ...  The style-based GAN architecture (StyleGAN) yields state-of-the-art results in data-driven unconditional generative image modeling.  ...  Acknowledgements We thank Ming-Yu Liu for an early review, Timo Viitanen for help with the public release, David Luebke for in-depth discussions and helpful comments, and Tero Kuosmanen for technical support  ... 
arXiv:1912.04958v2 fatcat:sruxmi3uuvdhnanr6iehk3gnpe

Analyzing and Improving the Image Quality of StyleGAN

Tero Karras, Samuli Laine, Miika Aittala, Janne Hellsten, Jaakko Lehtinen, Timo Aila
2020 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
Overall, our improved model redefines the state of the art in unconditional image modeling, both in terms of existing distribution quality metrics as well as perceived image quality.  ...  The style-based GAN architecture (StyleGAN) yields state-of-the-art results in data-driven unconditional generative image modeling.  ...  Acknowledgements We thank Ming-Yu Liu for an early review, Timo Viitanen for help with the public release, David Luebke for in-depth discussions and helpful comments, and Tero Kuosmanen for technical support  ... 
doi:10.1109/cvpr42600.2020.00813 dblp:conf/cvpr/KarrasLAHLA20 fatcat:5jlnui35argwriq4jbeqbrsz5u

Attributing and Detecting Fake Images Generated by Known GANs

Matthew Joslin, Shuang Hao
2020 2020 IEEE Security and Privacy Workshops (SPW)  
The quality of GAN-generated fake images has improved significantly, and recent GAN approaches, such as StyleGAN, achieve near indistinguishability from real images for the naked eye.  ...  In this paper, we analyze the frequency domain to attribute and detect fake images generated by a known GAN model.  ...  INTRODUCTION Since the introduction of Generative Adversarial Networks (GANs) which can be used to generate superficially realistic images, the quality of these synthetic images has progressively improved  ... 
doi:10.1109/spw50608.2020.00019 fatcat:5byl2rzdl5a7xfwmdngbgdd2l4

Automated Sewer Defects Detection Using Style-based Generative Adversarial Networks and Fine-tuned Well-known CNN Classifier

Zuxiang Situ, Shuai Teng, Hanlin Liu, Jinhua Luo, Qianqian Zhou
2021 IEEE Access  
Date of publication xxxx 00, 0000, date of current version xxxx 00, 0000.  ...  Improved strategy 2: StyleGANs implement the transfer learning [45] [46] [47] to enhance model performance in terms of image quality and training speed.  ...  The BigGAN achieved a great improvement in the quality of synthetic images, but it requires a large number of training samples and time.  ... 
doi:10.1109/access.2021.3073915 fatcat:gkybbs5xercrhovvsnxmey2zou

Detecting GAN generated errors [article]

Xiru Zhu, Fengdi Che, Tianzi Yang, Tzuyang Yu, David Meger, Gregory Dudek
2019 arXiv   pre-print
This is because the task of evaluating the quality of a generated image differs from deciding if an image is real or fake.  ...  We leverage Improved Wasserstein, BigGAN, and StyleGAN to show a ranking based on our metric correlates impressively with FID scores.  ...  This approach consists of analyzing generated images and finding the size of a dataset needed before duplicate samples appear.  ... 
arXiv:1912.00527v1 fatcat:qyzf7ix3mbfklackr2xrezpogu

State-of-the-Art in the Architecture, Methods and Applications of StyleGAN [article]

Amit H. Bermano and Rinon Gal and Yuval Alaluf and Ron Mokady and Yotam Nitzan and Omer Tov and Or Patashnik and Daniel Cohen-Or
2022 arXiv   pre-print
This state-of-the-art report covers the StyleGAN architecture, and the ways it has been employed since its conception, while also analyzing its severe limitations.  ...  Of these, StyleGAN offers a fascinating case study, owing to its remarkable visual quality and an ability to support a large array of downstream tasks.  ...  More specifically, StyleGAN [Karras et al. 2019 ] has reached the forefront of image synthesis, gaining recognition as the state-of-the-art generator for high-quality images and becoming the de-facto  ... 
arXiv:2202.14020v1 fatcat:qu3plbdnszdujcwxwq3zizysje

Improved StyleGAN Embedding: Where are the Good Latents? [article]

Peihao Zhu, Rameen Abdal, Yipeng Qin, John Femiani, Peter Wonka
2021 arXiv   pre-print
First, we introduce a new normalized space to analyze the diversity and the quality of the reconstructed latent codes.  ...  Second, we propose an improved embedding algorithm using a novel regularization method based on our analysis. Finally, we analyze the quality of different embedding algorithms.  ...  Our main contributions are: • the introduction of and + space for analyzing and regularizing StyleGAN embeddings. • a new regularizer for StyleGAN embedding that provides the best trade-off between reconstruction  ... 
arXiv:2012.09036v3 fatcat:adyjsrwehna77i77vu2bstrhc4

One Model to Reconstruct Them All: A Novel Way to Use the Stochastic Noise in StyleGAN [article]

Christian Bartz, Joseph Bethge, Haojin Yang, Christoph Meinel
2020 arXiv   pre-print
Different works have improved the limited understanding of the latent space of GANs by embedding images into specific GAN architectures to reconstruct the original images.  ...  Furthermore, we provide new insights about the significance and capabilities of noise inputs of the well-known StyleGAN architecture.  ...  Acknowledgement We wish to thank the Wildenstein Plattner Institute for supplying us with data and a research objective that ultimately led to the results published in this paper.  ... 
arXiv:2010.11113v1 fatcat:rwwygcfplfgypbrsfzfftqln5y

An Overview of Face Manipulation Detection

Xingwang Ju
2020 Journal of Cyber Security  
On this basis, the challenges for the research and the potential future directions are analyzed and discussed.  ...  Due to the power of editing tools, new types of fake faces are being created and synthesized, which has attracted great attention on social media.  ...  It generates two different qualities of image: 64 × 64 (low-quality images), and 128 × 128 (high-quality images).  ... 
doi:10.32604/jcs.2020.014310 fatcat:fcu67ro5pvcg7eputna6dlf5eu

Fine-Grained Control of Artistic Styles in Image Generation [article]

Xin Miao, Huayan Wang, Jun Fu, Jiayi Liu, Shen Wang, Zhenyu Liao
2021 arXiv   pre-print
Experiments show that our method not only precisely controls the fine-grained artistic style but also improves image quality over vanilla StyleGAN as measured by FID.  ...  Few works have been explored to capture the continuous spectrum of styles and apply it to a style generation task.  ...  It was proved to be able to significantly improve the quality of class conditional image generation.  ... 
arXiv:2110.10278v2 fatcat:fovdj7j4cfhs5iyf4ukletf7ta

Thermal Face Generation using StyleGAN

Gabriel Hermosilla, Diego-Ignacio Henriquez Tapia, Hector Allende-Cid, Gonzalo Farias, Esteban Vera
2021 IEEE Access  
In addition, a quantitative analysis will be carried out to analyze the quality of the images generated.  ...  Please observe in Figure 3 the evolution of the GAN in recent years and how the quality and resolution of face generation have improved.  ... 
doi:10.1109/access.2021.3085423 fatcat:2couovxccfhlxneh3kgojyispa

Transforming the Latent Space of StyleGAN for Real Face Editing [article]

Heyi Li, Jinlong Liu, Xinyu Zhang, Yunzhi Bai, Huayan Wang, Klaus Mueller
2022 arXiv   pre-print
Our modified StyleGAN maintains the state-of-the-art generation quality of the original StyleGAN with moderately better diversity.  ...  The gap between the W space and the W+ space demands an undesirable trade-off between reconstruction quality and editing quality.  ...  Acknowledgments We would like to show our gratitude to Jun Fu for advice on the design of figures; Jiayi Liu, Shen Wang, and Zhihang Li for paper reviews.  ... 
arXiv:2105.14230v2 fatcat:zmqmpuykyndn3hrckd6irsube4

StyleGAN-XL: Scaling StyleGAN to Large Diverse Datasets [article]

Axel Sauer, Katja Schwarz, Andreas Geiger
2022 arXiv   pre-print
Our final model, StyleGAN-XL, sets a new state-of-the-art on large-scale image synthesis and is the first to generate images at a resolution of 1024^2 at such a dataset scale.  ...  StyleGAN in particular sets new standards for generative modeling regarding image quality and controllability.  ...  ACKNOWLEDGMENTS We acknowledge the financial support by the BMWi in the project KI Delta Learning (project number 19A19013O). Andreas Geiger was supported by the ERC Starting Grant LEGO-3D (850533).  ... 
arXiv:2202.00273v2 fatcat:ecxudyms7bf77d7jwwxro362tm

Dermoscopy Image Classification based on StyleGAN and DenseNet201

Chen Zhao, Renjun Shuai, Li Ma, Wenjia Liu, Die Hu, Menglin Wu
2021 IEEE Access  
The proposed method improves the accuracy of skin lesion image classification, assists dermatologists in determining and diagnosing different types of skin lesions, and analyzes skin lesions at different  ...  The proposed framework redesigns the structure of style control and noise input in the original generator and reconstructs the discriminator to adjust the generator to efficiently synthesize high-quality  ...  ACKNOWLEDGMENT Chen Zhao contributed to the writing and editing of the article and the operation and editing of the code.  ... 
doi:10.1109/access.2021.3049600 fatcat:2mynogmtsrfbtl7m46fwjryamq

Semi-Supervised StyleGAN for Disentanglement Learning [article]

Weili Nie, Tero Karras, Animesh Garg, Shoubhik Debnath, Anjul Patney, Ankit B. Patel, Anima Anandkumar
2020 arXiv   pre-print
We investigate the impact of limited supervision and find that using only 0.25%~2.5% of labeled data is sufficient for good disentanglement on both synthetic and real datasets.  ...  Current disentanglement methods face several inherent limitations: difficulty with high-resolution images, primarily focusing on learning disentangled representations, and non-identifiability due to the  ...  We have also increased the network sizes of disentangled VAEs and InfoGAN-CR, but the improvement of image quality still cannot close the gap with Info-StyleGAN, as shown in Appendix B.4.  ... 
arXiv:2003.03461v3 fatcat:qjavkgece5evbku3rfon55mqr4
« Previous Showing results 1 — 15 out of 745 results