Machine Learning & Computational Design Footwear Explorations
The following explorations are the result of a collaboration with computational design researching ways to integrate data into the knitting process with AI, ML, and CD.
Footwear
Research Prompt: can you design a tunable, customizable performance upper that is engineered with zones of stiffness, stretchiness, and breathability using only two yarns, no intarsia zones, and a combination of two knit unique structures?
After a series of trials, I landed on two different knit structures, one that had zero stretch and one that was very stretchy and open, that were compatible in different combinations of lengths and variations. We ran Instron pull testing trials to verify these two structures had vastly different properties and were strong enough to withstand the performance requirements of a running shoe.
After confirming which two structures we were going to use, we generated ML images in Runway for graphic inspiration and combined these unique patterns with an upper stretch map data and N logo. These images were convered to 6 color bmp files and brought into knit software. I created 6 combinations of the two different knit structures and assigned each a color corresponding to the graphic.
The resulting knit uppers have unique, customizable uppers with distinct stretchy and stretch-free zones that correspond with performance footwear stretch map data.
