Researchers at the Massachusetts Institute of Technology (MIT) are creating AI-based software to help anyone make knitted garments, even those who have never knitted before in their lives.
The researchers at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have released two systems to make industrial knitting with machines as accessible as 3D printing.
“The impact of 3-D knitting has the potential to be even bigger than that of 3D printing,” said Prof. Jim McCann from the Carnegie Mellon Textiles Lab, as reported by Venture Beat. “Right now, design tools are holding the technology back, which is why this research is so important to the future.”
InverseKnit translates photos of knitted patterns into knitting instructions, whereas CADKnit helps users knit design patterns through images, computer-aided design software and photo design techniques.
To prepare the InverseKnit programme, the team needed to compile a data set of 17 different knitting instructions through Shima Seiki’s KnitPaint software along with matching images. The photos were generated by knitting a subset of 1,044 real-world patches and rendering the patches through KnitPaint’s pattern preview feature.
As a result, the AI algorithm was taught how to interpret the knitting patterns from the images, with the ultimate goal of creating machine-readable instructions just from seeing a picture of a knitted object.
“Current state-of-the-art computer vision techniques are data-hungry, and they need many examples to model the world effectively,” said Prof. McCann. “With InverseKnit, the team collected an immense data set of knit samples that, for the first time, enables modern computer vision techniques to be used to recognize and parse knitting patterns.”
According to the research team, InverseKnit generates accurate instructions 94 percent of the time. Although it currently works with just a sample size and only acrylic yarn, they hope to expand both in the future.
CADKnit was created with causal users and beginners in mind, allowing them to write their own programmes or visually manipulate the corresponding shapes and patterns to customise the design. It evens warns users when a pattern is “undesirable”
“As far as machines and knitting go, this type of system could change accessibility for people looking to be the designers of their own items,” said Alexandre Kaspar, computer scientist and head of the initiative. “We want to let casual users get access to machines without needing programming expertise, so they can reap the benefits of customisation by making use of machine learning for design and manufacturing.”
CADKnit was tested on casual users to create patterns for socks, hats, sweatpants, yoked shirts and scarfs. After using the software, most said that the programme was easy to use.
“Whether it’s for the everyday user who wants to mimic a friend’s beanie hat, or a subset of the public who might benefit from using this tool in a manufacturing setting, we’re aiming to make the process more accessible for personal customisation,” Kaspar explained, as reported by Engadget.
However, the software was incompatible with sweaters, was limited to one type of yarn and can currently only use basic patterns.
“3D printing took a while before people were comfortable enough to think they could do something with it,” Kaspar told TechCrunch. “It will be the same thing with what we do.”