Efficient and sustainable fashion design: AI-tool for creative creation

Purpose and goal:

This project addresses the mainstay of the fashion industry, high-fashion, specifically the early design phase characterized by uniqueness and creativity. The goal is to use generative AI, specifically generative image models, to promote human creativity and innovation as well as demonstrate and inspire the industry to apply AI. The project aims to develop a tool for the early design phase without compromising intellectual property rights, and instead capture a designer´s unique expression.

Expected results and effects:

The outcome of the project is to develop a tool for the design process for high-fashion in the fashion industry where generative image models (AI) are the engine. The tool is to be built in the form of a service that enables dissemination to the industry, where Sweden is of utmost importance. The project impact is: (1) increased degree of automation in the design process, (2) increased efficiency in the design process (time-to-market) and (3) reduced waste in the form of material consumption.

Approach and implementation:

The project is implemented in four phases; Learn, Ideate, Build and Launch, with weekly stand-ups and monthly project meetings. Requirements gathering, building datasets, training models, implementation and production take place iteratively and are continuously tested. The tool will be demonstrated in several rounds and the results will be disseminated in industry-specific contexts. The project is led by Substorm AB with Petra Fagerström AB as demanding party with industry knowledge.