Walking Through Shanshui: Generating Chinese Shanshui Paintings via Real-Time Tracking of Human Position

Aven Le Zhou; Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2019, pp. 0-0

Abstract


Shanshui is a traditional East Asian style of ink brush painting that depicts natural landscapes in a semi-abstract fashion. To create a Shanshui painting, ancient Chinese scholar-artists rely heavily on their travel experiences as well as their movements in natural spaces. In this paper, we propose an interactive system - "Walking Through Shanshui" - based on AI using Generative Adversarial Networks and various computer vision techniques. The system is an interactive art installation that helps bring the original experience of creating Shanshui to participants by tracking their movement through walking in a virtual space. It uses position tracking as input to generate Shanshui from participant's movements and to paint with a custom generative Sketch-to-Shanshui translation model. The system detects the participant's position in real-time and automatically traces it to generate a Shanshui painting instantly.

Related Material


[pdf]
[bibtex]
@InProceedings{Zhou_2019_ICCV,
author = {Le Zhou, Aven},
title = {Walking Through Shanshui: Generating Chinese Shanshui Paintings via Real-Time Tracking of Human Position},
booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops},
month = {Oct},
year = {2019}
}