Kyle McDonald and Lisa Kori use a Kinect and open source software to create custom jeans cheaply and efficiently.
How many pairs of jeans do you own that don't really fit you properly? Sure they used to fit when you bought them, and they fit nice for about two hours after they're fresh out of the cupboard after being washed but then, inevitably, they seem to grow ten sizes. That problem can probably be solved by spending a small fortune on designer brands—but who wants to do that? So how about creating your own custom pair, tailored to your body shape?
That's the idea behind Kyle McDonald and Lisa Kori Chung's Open Fit project. The project uses a Kinect camera to scan a person's body and then uses algorithms to generate tailored pants patterns. The software is open source and is available on GitHub for anyone to play around with. At the end of June the pair trialled their project at the Open Fit Lab, where the creation and sewing of these custom jeans became a performance.
McDonald's previous projects often feature open source software and his work often leads the field in R&D, opening it up for others to come and experiment, like he did with 3D scanning and face tracking. With this project he uses his 3D scanning knowledge and applies it to an idea that democratizes the world of custom-fit clothing.
McDonald reveals more about the project below.
The Creators Project: This project feels like a bit of a departure from your typical work. Can you tell us how your interest in fashion came about?
Kyle McDonald: During the NODE festival in Frankfurt this February, Lisa and I were talking about ideas for projects. My background is more in interactive art and computer vision, while she's spent a lot of time studying electronic art and the community, along with a healthy dose of hand crafted shoe making. She shared an idea she's had for years: creating a pop-up shop where people can have their measurements taken automatically, and receive a pair of custom fit jeans an hour later. Like a one hour photo center for jeans. I'd already talked with a lot of people about using 3D scanners to make measurements of bodies, so the part that got me really excited was when we started talking about the pattern generation.
A pair of OpenFit jeans
That's awesome. So how does the 3D scanning translate into generating a pair of custom jeans?
Before you sew a pair of jeans, you need to have a pattern to cut. And that pattern is derived from the body measurements using a complicated algorithm. While these algorithms can be found in tailoring manuals, most home sewers buy pre-made patterns drafted from very conventional designs. And the patterns and grading rules—the algorithms for sizing—used for mass-produced clothing are highly proprietary, built with software that can cost tens of thousands of dollars. With this software out of reach for small businesses and individual designers, many still use paper and go to pattern grading businesses that print the pattern in many sizes. But because pattern making is a series of geometric steps, the process, even for professional pattern makers and designers, still involves several passes of trial and error. Without the ability to quickly test what works and what doesn't, the pace of each design iteration can be really slow.
Whenever I hear about technology that is inaccessible to creative people, I get excited about crafting an open source entry point. So Lisa and I sat down and started recreating some of the "by the book" pattern making techniques using Processing, and from there Lisa implemented her own drafting algorithms. We tested those patterns, and then I created a system to take a 15 circumferential and lengthwise body measurements almost instantaneously with a Kinect using openFrameworks and techniques from computer vision.
Using a Kinect, a 3D scan of a person's body is created to generate the precise measurements of their body and achieve the perfect fit
The computer generates a pattern based on custom measurements extracted from the 3D scan
The pattern is then projected onto the wall and traced onto fabric, which is then cut and sewn to make jeans
So tell us a bit about how the OpenFit Lab came to life IRL and what the results were.
We thought the most fun way to test everything out would be to mashup the pop-up shop idea with a party, so we invited a bunch of friends to Pinion Gallery in Brooklyn and made the first six people custom fit jeans. The only catch was that they had to wear skin-tight bright green leggings in front of all their friends, so we made sure to provide some proper cheering and sufficient alcohol. Even though all the software is very beta, a few pairs fit really well, even with all the room for errors in measuring, cutting, and sewing.
The sewing team at Pinion Gallery
How do you plan on developing the project? Where do you hope to take it and what are your goals?
The next steps will be further refining the pattern making software, and making it easier for new coders to create their own dynamic patterns. We're also really interested in exploring a 3D approach to pattern making that uses open source tools to create hyper-accurate meshes, and flattening them out, instead of the usual 2D approach which can have strange artifacts (poor fit) in edge cases. Soon we'll be announcing our next OFL event, and we're excited to get more people involved making their own patterns, continuing this work towards more sharing of pattern making and clothing production knowledge.
I still find it amazing that such a basic collection of knowledge—the rules that govern the clothing that each of us wears every day—is so esoteric and traditionally constrained. Hopefully through this project we can start to open up those processes a bit.
All photos courtesy Kyle McDonald and Lisa Kori's Flickr pool.