Filters








6 Hits in 1.8 sec

CoCosNet v2: Full-Resolution Correspondence Learning for Image Translation [article]

Xingran Zhou, Bo Zhang, Ting Zhang, Pan Zhang, Jianmin Bao, Dong Chen, Zhongfei Zhang, Fang Wen
2021 arXiv   pre-print
We present the full-resolution correspondence learning for cross-domain images, which aids image translation.  ...  Experiments on diverse translation tasks show that CoCosNet v2 performs considerably better than state-of-the-art literature on producing high-resolution images.  ...  image translation. • To achieve that, we propose CoCosNet v2, a hierarchical GRU-assisted PatchMatch method, for efficient correspondence computation, which is simultaneously learned with image translation  ... 
arXiv:2012.02047v2 fatcat:pwfdpey2wvgipjcnruvqnihj3q

Bi-level Feature Alignment for Versatile Image Translation and Manipulation [article]

Fangneng Zhan, Yingchen Yu, Rongliang Wu, Jiahui Zhang, Kaiwen Cui, Aoran Xiao, Shijian Lu, Chunyan Miao
2022 arXiv   pre-print
This paper presents a versatile image translation and manipulation framework that achieves accurate semantic and style guidance in image generation by explicitly building a correspondence.  ...  Generative adversarial networks (GANs) have achieved great success in image translation and manipulation.  ...  Input Image Input Semantic Edited Semantic HIM SESAME CoCosNet CoCosNet v2 Ours Ground Truth Fig. 9.  ... 
arXiv:2107.03021v2 fatcat:jappnzfgmfdt7ce6mbmincboqq

DynaST: Dynamic Sparse Transformer for Exemplar-Guided Image Generation [article]

Songhua Liu, Jingwen Ye, Sucheng Ren, Xinchao Wang
2022 arXiv   pre-print
In addition, we introduce a unified training objective for DynaST, making it a versatile reference-based image translation framework for both supervised and unsupervised scenarios.  ...  One key challenge of exemplar-guided image generation lies in establishing fine-grained correspondences between input and guided images.  ...  [69] propose CoCosNet-v2 that is capable to learn full-resolution correspondence, the iterative global searching process by GRU makes a negative influence on the efficiency.  ... 
arXiv:2207.06124v2 fatcat:5faqwzu3mzfwjlvlezgpni7w2e

Image-to-Image Translation: Methods and Applications [article]

Yingxue Pang, Jianxin Lin, Tao Qin, Zhibo Chen
2021 arXiv   pre-print
Image-to-image translation (I2I) aims to transfer images from a source domain to a target domain while preserving the content representations.  ...  I2I has drawn increasing attention and made tremendous progress in recent years because of its wide range of applications in many computer vision and image processing problems, such as image synthesis,  ...  [128] therefore proposed a GRUassisted refinement module that applies PatchMatch in a hierarchy to first learn the full-resolution, 1024×1024, cross-domain semantic correspondence, namely CoCosNetv2  ... 
arXiv:2101.08629v2 fatcat:i6pywjwnvnhp3i7cmgza2slnle

Facial-Sketch Synthesis: A New Challenge [article]

Deng-Ping Fan, Ziling Huang, Peng Zheng, Hong Liu, Xuebin Qin, Luc Van Gool
2022 arXiv   pre-print
We first introduce a high-quality dataset for FSS, named FS2K, which consists of 2,104 image-sketch pairs spanning three types of sketch styles, image backgrounds, lighting conditions, skin colors, and  ...  35 image-to-sketch approaches.  ...  Zhou et al. further extended CoCosNet with full-resolution semantic correspondence learning [82] , with the main difference being the use of a regular and GRU-based propagation applied iteratively at  ... 
arXiv:2112.15439v5 fatcat:lr4dofk57ffenklivpadkqjw2e

CoGS: Controllable Generation and Search from Sketch and Style [article]

Cusuh Ham, Gemma Canet Tarres, Tu Bui, James Hays, Zhe Lin, John Collomosse
2022 arXiv   pre-print
We present CoGS, a novel method for the style-conditioned, sketch-driven synthesis of images.  ...  CoGS enables exploration of diverse appearance possibilities for a given sketched object, enabling decoupled control over the structure and the appearance of the output.  ...  CoCosNet v2 [66] is a patch-based method that uses examplar images from different domains, such as edge maps or semantic maps, to translate into highquality photorealistic images.  ... 
arXiv:2203.09554v2 fatcat:hfuvh3yiojbndcce6ej7efvo2m