Challenge Introduction

Portrait sketching is an art form that uses black, white, and gray to create an image of a person through the use of lines. Different artistic styles and techniques can convey various emotions and moods. However, traditional portrait sketching heavily relies on the personal drawing skills of the artist and requires significant time and effort for each drawing. Thanks to the development of image style transfer, AI-assisted portrait sketching can significantly reduce the cost of creating portraits. However, the most common method of generating portrait sketches is based on input photos, which may not be available in all situations.

To address the issue of generating portrait sketches without photos and enable ordinary people to quickly obtain high-quality portrait sketches that meet their requirements, we have organized a portrait sketching generation challenge.

Task: In this challenge, we provide a dataset that includes both normal facial lines and professional portrait sketches. Participants are required to train on this dataset, using rough facial line images as input to generate corresponding portrait sketches as output. The generated results are then to be uploaded to the competition platform and evaluated using two performance metrics to determine the final ranking.

Challenge Organization

Challenge Co-Chairs

Organizing Team

  • Mengsi Guo, Wuhan Textile University, Wuhan, China
  • Lei Bi, University of Sydney, Sydney, Australia
  • Ping Li, The Hong Kong Polytechnic University, Hong Kong, China
  • Jin Huang, Wuhan Textile University, Wuhan, China
  • Bin Sheng, Shanghai Jiao Tong University, China
  • Xinrong Hu, Wuhan Textile University, China
  • Qinya Li, Shanghai Jiao Tong University, China

Challenge Webpage


Important Dates(For paper author)

  • Submission Deadline: 12 June (extented to 19 June) 2023 , GMT 23:59
  • Notification of Paper Decisions: 13 July 2023
  • Camera-Ready: 5 August 2023

Challenge Contact

Mengsi Guo, Wuhan Textile University, Wuhan, China