Following the initial setup - here is a test for an early idea of integrating the wearer into the production of a bespoke garment. The process begins by capturing the wearer's signature and using it to generate an image of values that define the fabric of the dress through a custom GAN.
A PDG network was built to generate the training data for the GAN. The output image corresponds to the edge length distribution in the cloth - this metadata is applied via UVs. Now even by keeping the pattern the same, every new wearer can individualise and alter the drape via their signature. This was quite an exciting moment! A neat way to utilise both AI and concepts from vision. To be developed...
