A group of university researchers developed artificial intelligence (AI) that allows people to create their knitting patterns to make clothes all by themselves without training.
Researchers have long thought of having an option for people to make their clothes, as most people nowadays still buy clothes off the rack and without any choice but to wear those that are sold in the department store.
As engineers, a team of researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), led by computer scientist Alexandre Kaspar, thought precisely that — train an AI to create beautiful patterns so people can make clothes by themselves.
The project launched two different studies as part of their knit-it-yourself mission. The first of these papers is about their newly developed software, InverseKnit, which creates knitting patterns out of a photo of a garment or fabric. As the name suggests, InverseKnit inverses the process of knitting, developing the pattern by which the sample fabric was knitted from and translate these photos into knitting instructions.
The second of these interesting developments is CADKnit, which is basically a design software, where users can create their knitting designs. The software allows users to customize their knitting patterns, customizing the template, changing sizes, the final shape, and decorative details. CADKnit uses a combination of two-dimensional images, computer-aided design software, and photo design techniques to let users knit design templates.
How do they work?
The team behind the two software trained an AI algorithm to “teach” them to interpret two-dimensional knitting instructions from images so that they could translate them into machine-readable formats. The instructions triggered one of several attached knitting machine’s basic operations, like a knit (which pulls a loop of yarn through all current loops), a tuck (which stacks a new loop onto a needle), a miss (which skips a needle), or a transfer (which moves the needle’s content to the other bed).
The tests on both software produce accurate pattern translation is 94% of the time, according to researchers. However, for now, the algorithm only works for a small sample size (and only with acrylic yarn) and doesn’t explicitly model the pattern scale. Nonetheless, the researchers are optimistic that through time, they can develop the software that overcomes its current constraints and can be commercially distributed.
“Current state-of-the-art computer vision techniques are data-hungry, and they need many examples to model the world effectively,” said Prof. Jim McCann, who works on similar projects at the unrelated Carnegie Mellon Textiles Lab. “With InverseKnit, the team collected an immense dataset of knit samples that, for the first time, enables modern computer vision techniques to be used to recognize and parse knitting patterns.”
They could be as available to people as 3D printing
In the past few years, the development of 3D scanning has allowed people to make things on their own, allowing students to create models and prototypes, businesses to develop their own products, and anyone to create small items out of a 3D printer. However, up until today, making your own clothes, especially if you’re not a seamster and have never had any knowledge about knitting, is an ordeal that latest technology is yet to simplify.
That’s why the researchers are passionate about their project. They said that the products and designs developed in both InverseKnit and CADKnit could be fed to an already-commercially available knitting machine to make clothes. These commercial knitting machines have already been available to some knitters, but it takes a certain level of technical knowledge to design a pattern that will work for these machines. Now, MIT researchers have developed software that democratizes knitting.
The goal of both software, CADKnit and InverseKnit, is to make knitting as accessible to as many people as 3D printing is nowadays. As part of Kaspar’s ambition for the software when they are launched commercially, he envisions that “knitting as a service” for consumers who want to order customized garments. It can also enable clothing designers to spend less time learning how to write knitwear patterns for machines and reduce waste in the prototyping and manufacturing process. Another target audience for the software are hand-knitters who want to try a new way of working with yarn.”
“If you think about it like 3D printing, a lot of people have been using 3D printers or hacking 3D printers, so they are great potential users for our system because they can do that with knitting,” Kaspar added.