BézierSketch: A Generative Model for Scalable Vector Sketches

Ayan Das; Yongxin Yang; Timothy M Hospedales; Tao  Xiang; Yi-Zhe Song

doi:10.1007/978-3-030-58574-7_38

BézierSketch: A Generative Model for Scalable Vector Sketches

Ayan Das, Yongxin Yang, Timothy M Hospedales, Tao Xiang, Yi-Zhe Song

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

The study of neural generative models of human sketches is a fascinating contemporary modeling problem due to the links between sketch image generation and the human drawing process. The landmark SketchRNN provided breakthrough by sequentially generating sketches as a sequence of waypoints. However this leads to low-resolution image generation, and failure to model long sketches. In this paper we present BézierSketch, a novel generative model for fully vector sketches that are automatically scalable and high-resolution. To this end, we first introduce a novel inverse graphics approach to stroke embedding that trains an encoder to embed each stroke to its best fit Bézier curve. This enables us to treat sketches as short sequences of paramaterized strokes and thus train a recurrent sketch generator with greater capacity for longer sketches, while producing scalable high-resolution results. We report qualitative and quantitative results on the Quick, Draw! benchmark.

Original language	English
Title of host publication	Computer Vision – ECCV 2020
Subtitle of host publication	16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXVI
Publisher	Springer, Cham
Pages	632-647
Number of pages	19
ISBN (Electronic)	978-3-030-58574-7
ISBN (Print)	978-3-030-58573-0
DOIs	https://doi.org/10.1007/978-3-030-58574-7_38
Publication status	Published - 13 Nov 2020
Event	16th European Conference on Computer Vision - Virtual conference Duration: 23 Aug 2020 → 28 Aug 2020 https://eccv2020.eu/

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer, Cham
Volume	12371
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	16th European Conference on Computer Vision
Abbreviated title	ECCV 2020
City	Virtual conference
Period	23/08/20 → 28/08/20
Internet address	https://eccv2020.eu/

Keywords

Sketch generation
Scalable graphics
Bézier curve

Access to Document

10.1007/978-3-030-58574-7_38Licence: All Rights Reserved

BezierSketch_DAS_DOA02072020_AFVAccepted author manuscript, 2.59 MBLicence: All Rights Reserved

https://link.springer.com/chapter/10.1007/978-3-030-58574-7_38Licence: All Rights Reserved

Cite this

Das, A., Yang, Y., Hospedales, T. M., Xiang, T., & Song, Y-Z. (2020). BézierSketch: A Generative Model for Scalable Vector Sketches. In Computer Vision – ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXVI (pp. 632-647). (Lecture Notes in Computer Science ; Vol. 12371). Springer, Cham. https://doi.org/10.1007/978-3-030-58574-7_38

@inproceedings{82c7b75997614c1f81881ca537d6ef26,

title = "B{\'e}zierSketch: A Generative Model for Scalable Vector Sketches",

abstract = "The study of neural generative models of human sketches is a fascinating contemporary modeling problem due to the links between sketch image generation and the human drawing process. The landmark SketchRNN provided breakthrough by sequentially generating sketches as a sequence of waypoints. However this leads to low-resolution image generation, and failure to model long sketches. In this paper we present B{\'e}zierSketch, a novel generative model for fully vector sketches that are automatically scalable and high-resolution. To this end, we first introduce a novel inverse graphics approach to stroke embedding that trains an encoder to embed each stroke to its best fit B{\'e}zier curve. This enables us to treat sketches as short sequences of paramaterized strokes and thus train a recurrent sketch generator with greater capacity for longer sketches, while producing scalable high-resolution results. We report qualitative and quantitative results on the Quick, Draw! benchmark. ",

keywords = "Sketch generation, Scalable graphics, B{\'e}zier curve",

author = "Ayan Das and Yongxin Yang and Hospedales, {Timothy M} and Tao Xiang and Yi-Zhe Song",

year = "2020",

month = nov,

day = "13",

doi = "10.1007/978-3-030-58574-7_38",

language = "English",

isbn = "978-3-030-58573-0",

series = "Lecture Notes in Computer Science ",

publisher = "Springer, Cham",

pages = "632--647",

booktitle = "Computer Vision – ECCV 2020",

note = "16th European Conference on Computer Vision, ECCV 2020 ; Conference date: 23-08-2020 Through 28-08-2020",

url = "https://eccv2020.eu/",

}

Das, A, Yang, Y, Hospedales, TM, Xiang, T & Song, Y-Z 2020, BézierSketch: A Generative Model for Scalable Vector Sketches. in Computer Vision – ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXVI. Lecture Notes in Computer Science , vol. 12371, Springer, Cham, pp. 632-647, 16th European Conference on Computer Vision, Virtual conference, 23/08/20. https://doi.org/10.1007/978-3-030-58574-7_38

BézierSketch: A Generative Model for Scalable Vector Sketches. / Das, Ayan; Yang, Yongxin; Hospedales, Timothy M et al.

Computer Vision – ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXVI. Springer, Cham, 2020. p. 632-647 (Lecture Notes in Computer Science ; Vol. 12371).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - BézierSketch: A Generative Model for Scalable Vector Sketches

AU - Das, Ayan

AU - Yang, Yongxin

AU - Hospedales, Timothy M

AU - Xiang, Tao

AU - Song, Yi-Zhe

PY - 2020/11/13

Y1 - 2020/11/13

N2 - The study of neural generative models of human sketches is a fascinating contemporary modeling problem due to the links between sketch image generation and the human drawing process. The landmark SketchRNN provided breakthrough by sequentially generating sketches as a sequence of waypoints. However this leads to low-resolution image generation, and failure to model long sketches. In this paper we present BézierSketch, a novel generative model for fully vector sketches that are automatically scalable and high-resolution. To this end, we first introduce a novel inverse graphics approach to stroke embedding that trains an encoder to embed each stroke to its best fit Bézier curve. This enables us to treat sketches as short sequences of paramaterized strokes and thus train a recurrent sketch generator with greater capacity for longer sketches, while producing scalable high-resolution results. We report qualitative and quantitative results on the Quick, Draw! benchmark.

AB - The study of neural generative models of human sketches is a fascinating contemporary modeling problem due to the links between sketch image generation and the human drawing process. The landmark SketchRNN provided breakthrough by sequentially generating sketches as a sequence of waypoints. However this leads to low-resolution image generation, and failure to model long sketches. In this paper we present BézierSketch, a novel generative model for fully vector sketches that are automatically scalable and high-resolution. To this end, we first introduce a novel inverse graphics approach to stroke embedding that trains an encoder to embed each stroke to its best fit Bézier curve. This enables us to treat sketches as short sequences of paramaterized strokes and thus train a recurrent sketch generator with greater capacity for longer sketches, while producing scalable high-resolution results. We report qualitative and quantitative results on the Quick, Draw! benchmark.

KW - Sketch generation

KW - Scalable graphics

KW - Bézier curve

U2 - 10.1007/978-3-030-58574-7_38

DO - 10.1007/978-3-030-58574-7_38

M3 - Conference contribution

SN - 978-3-030-58573-0

T3 - Lecture Notes in Computer Science

SP - 632

EP - 647

BT - Computer Vision – ECCV 2020

PB - Springer, Cham

T2 - 16th European Conference on Computer Vision

Y2 - 23 August 2020 through 28 August 2020

ER -

BézierSketch: A Generative Model for Scalable Vector Sketches

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this