Upstyle
AI fashion creating tool
Project Type
23 spring Harvard Design Engineering Studio project
Group project:
Yiqi Yan (Product design & Prototype)
Wenbo Zhang (Research & Fashion)
My role
Product designer
UX / UI Design
Motion designer
Demo prototyper
Timeline
Spring 2023
2024 iF Finalist
This project is nominated by the iF Jury Chairperson @Karen Reuther, and will be directly participated in the final jury panel in 2024 iF Awards.
Context
Upstyle is an image-assisted text-to-image AI tool that allows you to create fashion pieces with more accuracy
The current AI generative tools, e.g., Midjourney are mostly biased and struggle to generate images with high accuracy. While they can be useful for creative brainstorming, they often off the trail for users with specific design objectives.
There are two main reasons for this. Firstly, Midjourney's AI model was trained on an extensive dataset of images, resulting in biased output. Additionally, users are typically non-professionals and may struggle to formulate accurately structured prompts for descriptions.
Therefore, Upstyles aims to address the accuracy issue in generating fashion pieces. In this project, we focus on the use case of 'Redesigning unworn garments' to demonstrate the image-to-image generation logic.