Introduction

Sketch Your Ancient Arts: Deep Learning on Chinese Ink-Wash Paintings

Sketch Your Ancient Arts

From AlphaGo to Midjourney, machine learning is no longer a “fancy concept” but something that indeed gets involved in our daily lives. However, the ever-changing technology prompts us to consider the boundary between Human and Artificial Intelligence (AI). What can or can’t AI do?

Withholding this question, we bring machine learning and traditional ink-washing paintings (i.e., Landscape paintings) together to see if any interesting aesthetic results would be generated. After intensive training, our model successfully produces ink-washing paintings that are highly akin to real artwork. So, what will be the conclusion of this result? It is refuting that arts represent human uniqueness? Or proving that AI is creative?

We do not think this is the case. 

Basically, producing similar paintings does not imply similar creative intentions as artists, i.e., subjective feelings and resonance with nature. Additionally, no matter how exquisite the paintings generated by AI are, the appreciation is based on humans’ aesthetic resonance that inheriting by generations, i.e., humans make creative decisions. Hence, we think the discussion of AI and Arts should not be limited to “threat and dying out”, but how to use AI to unleash human creativity.

In short, we believe that the improvement of the AI model in arts has two constructive applications. From the macro perspective, it may inspire new artistic elements or styles; from the micro perceptive, it empowers individuals like us to create, learn, and visualize our imaginations. 

Accordingly, this project aims to bridge between:

1) Traditional ink-washing arts and the latest technology;

2) Daily application and Machine Learning specialized in deep learning. 

We will provide an online interactive platform that allows you to

  • Witness the AI-generated ink-washing paintings inherited from the CUHK library and art museum collection
  • Explore the changes of lines and inks in landscape paintings. 

Acknowledgement

Art Museum, the Chinese University of Hong Kong | New Asia College Ch’ien Mu Library

Produced by CUHK Library | Eva So | Henry Ngan