Creative Sketch Generation

By Songwei Ge, Vedanuj Goswami, C. Lawrence Zitnick, Devi Parikh. 2020

93 citations recorded by the source corpus.

Sketching or doodling is a popular creative activity that people engage in. However, most existing work in automatic sketch understanding or generation has focused on sketches that are quite mundane. In this work, we introduce two datasets of creative sketches -- Creative Birds and Creative Creatures -- containing 10k sketches each along with part annotations. We propose DoodlerGAN -- a part-based Generative Adversarial Network (GAN) -- to generate unseen compositions of novel part appearances. Quantitative evaluations as well as human studies demonstrate that sketches generated by our approach are more creative and of higher quality than existing approaches. In fact, in Creative Birds, subjects prefer sketches generated by DoodlerGAN over those drawn by humans! Our code can be found at this https URL and a demo can be found at this http URL.

Subfields: Computational Mechanics, Computer Graphics and Computer-Aided Design, Control and Systems Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

1. Tao A. Xiang
2. Timothy M. Hospedales
3. Yi-Zhe Song
4. Sang-Min Park
5. Young-Gab Kim
6. Fei Sha
7. Soravit Changpinyo
8. Wei-Lun Chao

In Computer Graphics and Computer-Aided Design

1. Tao A. Xiang
2. Timothy M. Hospedales
3. Yi-Zhe Song
4. Pinaki Nath Chowdhury
5. Sang-Min Park
6. Peng Xu
7. Ayan Kumar Bhunia
8. Aneeshan Sain

In Control and Systems Engineering

1. Tao A. Xiang
2. Timothy M. Hospedales
3. Yi-Zhe Song
4. Pinaki Nath Chowdhury
5. Peng Xu
6. Qiyue Yin
7. Ayan Kumar Bhunia
8. Aneeshan Sain

Citations in the corpus

• Doodle It Yourself: Class Incremental Learning by Drawing a Few Sketches (2022, background): “There is of course also the added benefit of injecting creativity to the classifier by sketching something off the user’s imagination [16], e.g., a “flying cow”?”
• Deep Learning for Free-Hand Sketch: A Survey (2020, background): “Sketch generation [5], [56], [126], [127], [128], [129], [130], [131] has grown rapidly in recent years as deep learningbased approaches easily outperform earlier classic sketch generators [132], [133].”
• A Metaverse: Taxonomy, Components, Applications, and Open Challenges (2022, background): “[265] proposed DoodlerGAN, a generative partbased GAN (Generative Adversarial Network) that gener-”
• Partially Does It: Towards Scene-Level FG-SBIR with Partial Input (2022, background): “The field has predominately focused on objectlevel sketches to date [22, 26, 60], studying its abstractness [33], creativeness [26], and applications such as image retrieval [9, 79], and 3D synthesis/editing [30].”