ShoeGenAI: A Creativity Support Tool Bridging Design Intention and Feasibility in Shoe Design
summary
ShoeGenAI builds three models—a generator fine-tuned on real sneaker images and human descriptors, a combiner that crops soles with SAM then reconstructs with an encoder plus Stable Diffusion, and a Pix2Pix-based editor for localized fixes—and reports a 20-designer user study showing outputs are closer to intent and more manufacturable than DALL‑E 3. Template-driven prompts and hybrid part recombination cut repetitive Photoshop work; professionals prioritize feasibility while novices use templates to explore, and the team confirms factories can produce the generated sneaker designs.
AI-written summary · about
chapters (9)
- 0:00Introduction to ShoeGenAI
- 0:48Design process challenges
- 2:56GenAI limitations in shoe design
- 3:59ShoeGenAI system goals
- 4:20Core modules of ShoeGenAI
- 6:10User study and evaluation
- 8:17Professionals vs novices feedback
- 9:00Future work and applications
- 9:44Q&A session
related talks
20mGenerative Design
53mDesign Fashion with AI
29mTransformation in a New Era of AI
25mWalking in my Shoes - Defining a New Era in Footwear
25mArtificial intelligence & design: imagining the future of UX
21mOpportunities of Generative AI
39mAI & Creativity: A Designer's Journey Into the Future
19mAI-Powered Startup Toolkit for a Product Designer