Arbitrary style transfer using neurally-guided patch-based synthesis

Texler, Ondřej; Futschik, David; Fišer, Jakub; Lukáč, Michal; Lu, Jingwan; Shechtman, Eli; Sýkora, Daniel

doi:10.1016/j.cag.2020.01.002

Cited by 31 publications

(34 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Both the computation time for guiding channels and patch‐based synthesis and the memory footprint can further be reduced by using NNF upscaling method of Texler et al [TFF*20]. When NNF is upscaled two times, the resulting quality is still acceptable while the computational overhead is reduced to roughly 4 minutes and the size of 90 NNF samples is only 5MB.…”

Section: Resultsmentioning

confidence: 99%

“…Finally, in the real‐time rendering phase, the user browses to an arbitrary interaction state j ∊ I, and at that location, pre‐computed NNF s of nearby states x a , … x d are combined to produce NNF j from which the final target image T j is reconstructed using voting operation. Alternatively, the NNF upsampling technique of Texler et al [TFF*20] can be used to increase the resolution of the output image. See the text for further details.…”

Section: Our Approachmentioning

confidence: 99%

See 1 more Smart Citation

StyleProp: Real‐time Example‐based Stylization of 3D Models

Hauptfleisch

Texler

et al. 2020

Computer Graphics Forum

View full text Add to dashboard Cite

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Our Approachmentioning

confidence: 99%

StyleProp: Real‐time Example‐based Stylization of 3D Models

Hauptfleisch

Texler

et al. 2020

Computer Graphics Forum

View full text Add to dashboard Cite

show abstract

“…These techniques are limited to work at lower resolutions. Texler et al [2020] proposed two-stage process to improve the visual quality of neuralbased texture. They generate neural-based results, but use them as guidance for subsequent patch-based synthesis.…”

Section: Deep Learning Approach For Line Drawingsmentioning

confidence: 99%

Neurally-Guided Texturing for Garment Line Drawings

Hashimoto

Fukusato

Igarashi

2020

SIGGRAPH Asia 2020 Technical Communications

View full text Add to dashboard Cite

Adding texture to a line drawing is an important process in the production of comics and illustrations. Garment drawings especially often have large deformations with self-occlusions, so deforming patterns is essential for representing realistic garment designs. However, it is currently done manually and requires a significant amount of effort by experts. A possible approach is to infer 3D surface geometry and then apply texture to all 3D surfaces, but it is difficult to represent deep creases using this approach. In this paper, we introduce a "neurally-guided" optimization system for automatically deforming and directly mapping 2D texture patterns to 2D line drawings, bypassing 3D geometry. First, we build a deep neural network to estimate local transformation matrices of texture patterns, called neural-guidance, from line drawings. Second, we build a 2D triangle mesh for the garment and deforms the mesh to obtain the texture coordinates by integrating the local transformations. Our algorithm is effective and easy to integrate into existing drawing systems. We provide several examples to demonstrate the efficiency of our proposed system over previous methods and illustrate the versatility of our method. CCS CONCEPTS • Computing methodologies → Texturing; Shape modeling; Artificial intelligence.

show abstract

“…Our Journal will also be affected by these changes and we are working with many top venues to help in providing both quality control and speedy publication via Special and Technical Sections. This issue features papers from three special sections: 3DOR 2019 [1] , Expressive 2019 [2][3] and Graphics Interface 2019 [4] . Please do browse the Foreword to the Special Sections on Expressive 2019 and Graphics Interface that are closed in this issue.…”

Section: Editorial Notementioning

confidence: 99%

“…Our Facebook page 2 already reaches 2100 + people in the community and growing at a fast clip. A twitter feed 3 is another outlet that is experiencing a significant increase in following. In addition to these, we manage an Instagram page 4 and a YouTube 5 channel to boost the visibility of quality work as it appears on ScienceDirect.…”

Section: Online Presencementioning

confidence: 99%

Editorial Note

Jorge¹

2020

Computers & Graphics

View full text Add to dashboard Cite

Arbitrary style transfer using neurally-guided patch-based synthesis

Cited by 31 publications

References 14 publications

StyleProp: Real‐time Example‐based Stylization of 3D Models

StyleProp: Real‐time Example‐based Stylization of 3D Models

Neurally-Guided Texturing for Garment Line Drawings

Editorial Note

Contact Info

Product

Resources

About