Workshop

Overview

Sketching is one of the most powerful and versatile tools humans use to communicate their ideas. It has been used from ancient times to today, comes naturally to children, and transcends language barriers. The increasing prevalence of mobile computing devices with touchscreen interfaces has made it easier than ever to create and share sketches, as well as design popular games to collect large-scale sketch datasets (e.g., Quick, Draw!). The challenges posed by sketch understanding and generation have also helped to drive progress on fundamental problems in machine learning, including advances in the use of GANs (generative adversarial networks) and GNNs (graph neural network), as well as advances in meta-learning and self-supervised learning methods. These developments are reflected in the increasing number of research publications in top computer-vision venues that use sketching-related tasks or datasets, as well as the fact that as of 2021, the word “sketch” now appears as a separate submission theme in the CMT submission system at CVPR.

The goal of our second annual SketchDL workshop is to bring together perspectives from multiple discplines (e.g., computer vision, computer graphics, human computer interaction, deep learning, machine learning, cognitive science) to highlight major open questions and to identify collaboration opportunities to address outstanding challenges in this domain.

Program

08:30 - 08:35 . opening remarks (Room 238, local time, New Orleans, Louisiana)

08:35 - 09:30 . keynote by Azam Bastanfard

09:30 - 09:40 . oral paper presentation: "The Role of Shape for Domain Generalization on Sparsely-Textured Images"

09:40 - 09:50 . oral paper presentation: "SSR-GNNs: Stroke-based Sketch Representation with Graph Neural Networks"

09:50 - 10:00 . oral paper presentation: "Constellations: A novel dataset for studying iterative inference in humans and AI"

10:00 - 10:10 . oral paper presentation: "Leveraging Unlabeled Data for Sketch-based Understanding"

10:10 - 10:20 . oral paper presentation: "Signature Detection, Restoration, and Verification: A Novel Chinese Document Signature Forgery Detection Benchmark"

10:20 - 10:40 . break & poster session (Poster session is only for in-person audience.)

10:40 - 11:35 . keynote by Bria Long

11:35 - 12:30 . keynote by Jonas Jongejan

12:30 - 12:55 . free discussion for open problem

12:55 - 13:00 . closing remark

Call for Paper

(CFP poster can be downloaded via this link)

Topics of Interests

This workshop encourages novel and creative deep learning works for all forms of drawings, including free-hand sketch, professional (forensic) facial sketch, professional pencil sketch, professional landscape sketch, cartoon/manga, well-drawn 3D sketch, etc.

Topics of interests by this workshop include, but are not limited to:

sketch-related recognition, retrieval, and generation
novel algorithms for multi-modal visual understanding
novel datasets and benchmarks to evaluate robust multi-modal visual understanding
self-supervised/un-supervised/weakly-supervised representation learning methods for sketch
transfer learning, meta learning, zero-shot and few-shot learning for sketch
human-AI creative collaboration using sketches
conceptual blending and synthesis in sketches
novel art generation algorithms
evaluation protocols for generative art

Important Dates

Paper Submission Deadline . Apr. 3, 2022

Notification of Acceptance . Apr. 17, 2022

Camera-ready Due . Apr. 20, 2022

Submission and Review

All submissions will be handled electronically via the workshop’s CMT Website:
https://cmt3.research.microsoft.com/SketchDL2022.

All submissions will undergo standard double-blind peer-review.

Length, format, and template should follow the CVPR 2022 Submission Guidelines.

CVPR 2022, The 2nd Workshop on

Room 238. June 20th, 2022

(Link to 1st SketchDL workshop@CVPR2021)

Speakers

Organizers