Making Robots Draw A Vivid Portrait In Two Minutes

Gao, Fei; Zhu, Jingjie; Yu, Zeyuan; Li, Peng; Wang, Tao

doi:10.1109/iros45743.2020.9340940

Cited by 20 publications

(11 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2. First, we map both facial photos and sketches to line-drawings by using AiSketcher [20]: F : X /Y → Z. In this way, we obtain pseudo paired samples: {(z i , y i )} n i=1 with z i = F (y i ).…”

Section: Methodsmentioning

confidence: 99%

“…First, we use AiSketcher [20] as our line-drawing synthesizer, due to its remarkable performance for transferring multimodal images to line-drawings. AiSketcher is an extension of AdaIN [21] with a self-consistency loss and compositional sparse loss.…”

Section: Line-drawing Synthesismentioning

confidence: 99%

“…AiSketcher is an extension of AdaIN [21] with a self-consistency loss and compositional sparse loss. We reimplement AiSketcher exactly following [20], and use no paired data for training. In the testing stage, we apply the learned AiSketcher model F to facial photos or sketches, and obtain the corresponding line-drawings.…”

Section: Line-drawing Synthesismentioning

confidence: 99%

“…the line-drawings. To this end, we adopt a NST method, termed AiSketcher [20], and denotes it by F : X /Y → Z. As a result, we obtain pseudo paired data (Z, Y), and can learn the mapping from the line-drawing domain to the sketch domain, i.e.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Bridging Unpaired Facial Photos And Sketches By Line-drawings

Shang

Gao

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

In this paper, we propose a novel method to learn face sketch synthesis models by using unpaired data. Our main idea is bridging the photo domain X and the sketch domain Y by using the line-drawing domain Z. Specially, we map both photos and sketches to line-drawings by using a neural style transfer method, i.e. F : X /Y → Z. Consequently, we obtain pseudo paired data (Z, Y), and can learn the mapping G : Z → Y in a supervised learning manner. In the inference stage, given a facial photo, we can first transfer it to a line-drawing and then to a sketch by G • F . Additionally, we propose a novel stroke loss for generating different types of strokes. Our method, termed sRender, accords well with human artists' rendering process. Experimental results demonstrate that sRender can generate multi-style sketches, and significantly outperforms existing unpaired image-toimage translation methods.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Line-drawing Synthesismentioning

confidence: 99%

Section: Line-drawing Synthesismentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Bridging Unpaired Facial Photos And Sketches By Line-drawings

Shang

Gao

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…In [10], the authors presented a Cartesian robot for painting artworks by interactive segmentation of the areas to be painted and the interactive design of the orientation of the brush strokes. Furthermore, in [11], the authors adopted a machine learning approach for realizing brushstrokes as human artists, whereas in [12], a drawing robot, which can automatically transfer a facial picture to a vivid portrait and then draw it on paper within two minutes on average, was presented. More recently, in [13], a humanoid painting robot was shown, which draws facial features extracted robustly using deep learning techniques.…”

Section: Introductionmentioning

confidence: 99%

Robotic Sponge and Watercolor Painting Based on Image-Processing and Contour-Filling Algorithms

et al. 2022

View full text Add to dashboard Cite

In this paper, the implementation of a robotic painting system using a sponge and the watercolor painting technique is presented. A collection of tools for calibration and sponge support operations was designed and built. A contour-filling algorithm was developed, which defines the sponge positions and orientations in order to color the contour of a generic image. Finally, the proposed robotic system was employed to realize a painting combining etching and watercolor techniques. To the best of our knowledge, this is the first example of robotic painting that uses the watercolor technique and a sponge as the painting media.

show abstract