Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

DesignBridge: Bridging Designer Expertise and User Preferences through AI-Enhanced Co-Design for Fashion

Yuheng Shao, Yuansong Xu, Yifan Jin, Shuhao Zhang, Wenxin Gu, Quan Li

TL;DR

DesignBridge addresses the persistent gap between designer expertise and user preferences in fashion co-design by introducing a three-stage, AI-assisted workflow and dual interfaces for designers and users. The system combines an explicit design-space with intuitive preference elicitation (brush-based, scene-aware try-ons) and a preference-integrated generative loop that fine-tunes outputs based on consensus-driven signals. A formative study informs design goals and a nine-dimension garment space, while technical and user studies demonstrate that DesignBridge improves preference capture, analysis, and design quality relative to a baseline. The work advances interactive human-AI co-design by providing a scalable, interpretable framework that couples expert judgment with diverse user inputs to produce more acceptable, personalized fashion designs, with potential applicability across related creative domains.

Abstract

Effective collaboration between designers and users is important for fashion design, which can increase the user acceptance of fashion products and thereby create value. However, it remains an enduring challenge, as traditional designer-centric approaches restrict meaningful user participation, while user-driven methods demand design proficiency, often marginalizing professional creative judgment. Current co-design practices, including workshops and AI-assisted frameworks, struggle with low user engagement, inefficient preference collection, and difficulties in balancing user feedback with design considerations. To address these challenges, we conducted a formative study with designers and users experienced in co-design (N=7), identifying critical challenges for current collaboration between designers and users in the co-design process, and their requirements. Informed by these insights, we introduce DesignBridge, a multi-platform AI-enhanced interactive system that bridges designer expertise and user preferences through three stages: (1) Initial Design Framing, where designers define initial concepts. (2) Preference Expression Collection, where users intuitively articulate preferences via interactive tools. (3) Preference-Integrated Design, where designers use AI-assisted analytics to integrate feedback into cohesive designs. A user study demonstrates that DesignBridge significantly enhances user preference collection and analysis, enabling designers to integrate diverse preferences with professional expertise.

DesignBridge: Bridging Designer Expertise and User Preferences through AI-Enhanced Co-Design for Fashion

TL;DR

DesignBridge addresses the persistent gap between designer expertise and user preferences in fashion co-design by introducing a three-stage, AI-assisted workflow and dual interfaces for designers and users. The system combines an explicit design-space with intuitive preference elicitation (brush-based, scene-aware try-ons) and a preference-integrated generative loop that fine-tunes outputs based on consensus-driven signals. A formative study informs design goals and a nine-dimension garment space, while technical and user studies demonstrate that DesignBridge improves preference capture, analysis, and design quality relative to a baseline. The work advances interactive human-AI co-design by providing a scalable, interpretable framework that couples expert judgment with diverse user inputs to produce more acceptable, personalized fashion designs, with potential applicability across related creative domains.

Abstract

Effective collaboration between designers and users is important for fashion design, which can increase the user acceptance of fashion products and thereby create value. However, it remains an enduring challenge, as traditional designer-centric approaches restrict meaningful user participation, while user-driven methods demand design proficiency, often marginalizing professional creative judgment. Current co-design practices, including workshops and AI-assisted frameworks, struggle with low user engagement, inefficient preference collection, and difficulties in balancing user feedback with design considerations. To address these challenges, we conducted a formative study with designers and users experienced in co-design (N=7), identifying critical challenges for current collaboration between designers and users in the co-design process, and their requirements. Informed by these insights, we introduce DesignBridge, a multi-platform AI-enhanced interactive system that bridges designer expertise and user preferences through three stages: (1) Initial Design Framing, where designers define initial concepts. (2) Preference Expression Collection, where users intuitively articulate preferences via interactive tools. (3) Preference-Integrated Design, where designers use AI-assisted analytics to integrate feedback into cohesive designs. A user study demonstrates that DesignBridge significantly enhances user preference collection and analysis, enabling designers to integrate diverse preferences with professional expertise.
Paper Structure (70 sections, 3 equations, 10 figures, 10 tables)

This paper contains 70 sections, 3 equations, 10 figures, 10 tables.

Table of Contents

  1. INTRODUCTION
  2. BACKGROUND AND RELATED WORK
  3. Interaction and Collaboration between Designer and Users
  4. Preference Elicitation and Synthesis
  5. Generative Tools in Fashion Design
  6. FORMATIVE STUDY
  7. Methodology
  8. Findings
  9. C1. Ineffective Preference Collection Stemming from Low Engagement and Suboptimal Interaction
  10. C2. Coarse-Grained Preference Capture via Qualitative Methods
  11. C3. Demand for Multidimensional Feedback Integration in Iterative Design
  12. C4. Balancing Design Intent with User Preferences
  13. Design Goals
  14. Garment Design Space
  15. DESIGNBRIDGE
  16. ...and 55 more sections

Figures (10)

  • Figure 1: The entire process of the identification of garment design space, including collecting design image with corresponding description, conduct clustering analysis incorporating expert knowledge and research, and identify the definition of garment design space.
  • Figure 2: The workflow of DesignBridge is structured into three stages. In Stage One, designers provide initial inputs to identify a suitable design framework and generate corresponding images. In Stage Two, users evaluate the designs by viewing them both on a virtual model within a contextual scene and against a neutral background, expressing preferences at both the local and global levels. In Stage Three, designers analyze the collected user feedback, synthesize attributes across different dimensions of the design space, and iteratively refine the designs based on predicted user responses.
  • Figure 3: The Designer Interface of DesignBridge consists of four views: (A) The Framing Panel View receives the designer’s initial input to construct a preliminary design framework. (B) The Design Library View builds a design image database based on the design framework. (C) The Design Palette View integrates user feedback to assist designers in constructing the design. (D) The Informed Generation View presents the generated design images and the predicted user preference feedback of the current design.
  • Figure 4: The User Interface of DesignBridge consists of two views: (A) The Information View receives the user’s information to construct a virtual model for design try-on in the given scene. (B) The Interaction View enables users to express preferences for each design.
  • Figure 5: The detailed explanation of the tree structure and its interactions for an attribute. Designers can click a garment node in the first layer to view detailed user feedback in the second layer, showing interactions and comments. Designs with an equal or lower like-to-dislike ratio are classified as "dislike" and displayed above the root node.
  • ...and 5 more figures