Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

InkIdeator: Supporting Chinese-Style Visual Design Ideation via AI-Infused Exploration of Chinese Paintings

Shiwei Wu, Ziyao Gao, Zhendong He, Zongtan He, Zhupeng Huang, Xia Chen, Wei Zeng, Xiaojuan Ma, Zhenhui Peng

TL;DR

InkIdeator addresses the challenge of ideating Chinese-style visuals by pairing a formative study with a large, AI-annotated design space derived from 16,315 Chinese paintings. It couples a four-panel interface—Symbol Association, Image Library, Moodboard, and Image Generation—with a GenAI-driven pipeline to suggest symbols, analyze design dimensions, and generate visuals, enabling structured exploration and rapid concept visualization. A within-subject study (N=12) shows InkIdeator improves organized reference exploration, element extraction, and idea-to-visual translation, while expert painters demonstrate its potential for broader culture-related creative tasks. The work highlights the dual role of GenAIs in culture preservation, proposes design guidelines for culture-centric tools, and outlines avenues for future refinement, including knowledge-base augmentation and human-in-the-loop safeguards.

Abstract

Visual designers often seek inspiration from Chinese paintings when tasked with creating Chinese-style illustrations, posters, etc. Our formative study (N=10) reveals that during ideation, designers learn the cultural symbols, emotions, compositions, and styles in Chinese paintings but face challenges in searching, analyzing, and integrating these dimensions. This paper leverages multi-modal large models to annotate the value of each dimension in 16,315 Chinese paintings, built on which we propose InkIdeator, an ideation support system for Chinese-style visual designs. InkIdeator suggests cultural symbols associated with the task theme, provides dimensional keywords to help analyze Chinese paintings, and generates visual examples integrating user-selected keywords. Our within-subjects study (N=12) using a baseline system without extracted dimensional keywords, along with two extended use cases by Chinese painters, indicates InkIdeator's effectiveness in creative ideation support, helping users efficiently explore cultural dimensions in Chinese paintings and visualize their ideas. We discuss implications for supporting culture-related visual design ideation with generative AI.

InkIdeator: Supporting Chinese-Style Visual Design Ideation via AI-Infused Exploration of Chinese Paintings

TL;DR

InkIdeator addresses the challenge of ideating Chinese-style visuals by pairing a formative study with a large, AI-annotated design space derived from 16,315 Chinese paintings. It couples a four-panel interface—Symbol Association, Image Library, Moodboard, and Image Generation—with a GenAI-driven pipeline to suggest symbols, analyze design dimensions, and generate visuals, enabling structured exploration and rapid concept visualization. A within-subject study (N=12) shows InkIdeator improves organized reference exploration, element extraction, and idea-to-visual translation, while expert painters demonstrate its potential for broader culture-related creative tasks. The work highlights the dual role of GenAIs in culture preservation, proposes design guidelines for culture-centric tools, and outlines avenues for future refinement, including knowledge-base augmentation and human-in-the-loop safeguards.

Abstract

Visual designers often seek inspiration from Chinese paintings when tasked with creating Chinese-style illustrations, posters, etc. Our formative study (N=10) reveals that during ideation, designers learn the cultural symbols, emotions, compositions, and styles in Chinese paintings but face challenges in searching, analyzing, and integrating these dimensions. This paper leverages multi-modal large models to annotate the value of each dimension in 16,315 Chinese paintings, built on which we propose InkIdeator, an ideation support system for Chinese-style visual designs. InkIdeator suggests cultural symbols associated with the task theme, provides dimensional keywords to help analyze Chinese paintings, and generates visual examples integrating user-selected keywords. Our within-subjects study (N=12) using a baseline system without extracted dimensional keywords, along with two extended use cases by Chinese painters, indicates InkIdeator's effectiveness in creative ideation support, helping users efficiently explore cultural dimensions in Chinese paintings and visualize their ideas. We discuss implications for supporting culture-related visual design ideation with generative AI.
Paper Structure (50 sections, 15 figures, 4 tables)

This paper contains 50 sections, 15 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Chinese-Style Visual Design and Its Support Tools
  4. Example-Based Ideation Support
  5. Generative AI for Creative Support
  6. Formative study
  7. Participants and Procedure
  8. What Designers Learn from Chinese Paintings for Visual Design Ideation
  9. Perceptions towards GenAI Tools
  10. Challenges and Design Goals
  11. Designers would like to search for Chinese paintings related to the task but found it inefficient due to vague initial ideas and scattered resources.
  12. Designers would like to analyze the domain-specific dimensions of Chinese paintings but often found it difficult due to lack of domain knowledge.
  13. Designers would like to visualize their ideas, but it is time-consuming to manually sketch them and non-trivial to draft prompts to text-to-image generation models.
  14. Mining the Design Space in Chinese Paintings
  15. Data Collection
  16. ...and 35 more sections

Figures (15)

  • Figure 1: The interface of InkIdeator including: The Symbols Association Panel (A) recommends and explains cultural symbols based on the input design theme. The Image Library (B) allows searching examples from both our annotated Chinese painting dataset and online sources. The Moodboard Canvas (C) provides a space to arrange and analyze images by cultural symbols, emotions, compositions, and styles to inspire association. The Image Generation Panel (D) applies the explored keywords to generate sketches. This interface is originally in Chinese and translated in English for presentations in this paper.
  • Figure 2: System overview of InkIdeator. The system supports three iterative stages of Chinese-style visual design ideation—Search Reference, Analyze Reference, and Visualize Ideas—each addressing key challenges. This is achieved through core components, including the Symbol Association Panel, Image Library, Moodboard Canvas, and Image Generation Panel, connected to a Chinese Painting Dataset and powered by GPT and MidJourney for content generation.
  • Figure 3: Prompt template and output example of analyzing Chinese paintings using the multimodal large language model (MLLM).The left displays the Chinese painting being analyzed, the middle panel shows the prompt structure ( i.e., Role Play, Dimension Analysis, Chinese Painting Knowledge Injection, and Response Format) with example, and the right panel presents the structured output analyzed by MLLM.
  • Figure 4: Chinese painting examples that incorporate diverse cultural symbols, express different emotions, and reflect a range of artistic techniques (\ref{['tab:designspace']}).
  • Figure 5: Interaction with InkIdeator in user scenario. Steps 1–3: Diverge on a symbol under the environmental protection theme and ground it in the cultural context to launch ideation, then search example images to refine it. Steps 4–5: Analyze references from multiple perspectives of symbol, emotion, composition, and style, to promote the remaining search. Steps 6–10: Apply explored keywords to iteratively generate sketches.
  • ...and 10 more figures