MIT researchers with AI creates Knitting software

via

If you are familiar with MIT, you know how crazy the researchers over their goes. Yes, they go to next level. But sometimes they just redefine lives. The researchers at MIT came up with an weird idea of using AI for knitting. Not only they came up, technically MIT researchers with AI creates Knitting software. It looks as normal as it should, rather it can do incredible. The newly created version identifies patterns and combines them to create new ones. Now people who haven’t touched knitting. At all.

MIT researchers with AI creates Knitting software

Artificial Intelligence is an incredible thing to work with. Teach it, train it. And it will be able to replicate you with outmost perfection and to your worst nightmare. Also, Check this, AI is helping Hollywood to decide what movies to make.

But using it to the good, makes us happy. Let us keep it to that. Meanwhile, the researchers namely created two programs. One is the InverseKnit and the other is CADKnit.

What is InverseKnit?

via

This one specializes in creating patterns. It does by looking at photos of knitted items, for knitting software. InverseKnit makes use of advanced pattern recognition. Using so, it creates its own mixed patterns. The main part of the AI system is in this tool. A deep neural network trained with lots of patterns, to identify them. The tool is accurate to an error percentage of only 6%, which is incredible.

As it is commercial grade, with a few tweaking and tuning might be able to fix it up. The main work of the software is to identify the pattern, take it as an input. Then develop the newer patterns using the same data but in different colors and shapes. And it did with excellence.

What is CADKnit?

via

CADKnit does the work of the actual craft after InverseKnit. It allows the user to design, tune and customize the pattern templates. Even allows controls for size, dimensions, final touch and sizes along with extra decorative details. It combines the technique of Computer-Aided Drafting with photo editing tools for the templates. The software tool was handed over to newcomers, to the knitting field. While the tool was helpful, the users created incredible designs, in a short period of time.

Applications of knitting software

The software can be used to design the proper templates as liked and needed by the user. Then the templates can be fed into a knitting machine. The machine will do and knit, as it is instructed by the patterns. This can be of great advantage and efficient way of knitting. As traditional knitting and processes used in industries create a lot of wastes.

Commercial applications

via

The two software can deliver custom designs to online users right away. As they can access an online tool of both of the software, based on the same architecture. Or even an application that can get designed based on the tools. Users can download it, create their own designs and send it to the knitting industry. In this way, the customers can get their own customized design handicrafts. All without much hassles and worries of un-fittings.

Also for hand knitters, they can work with the machine with little studies on the technics. Yet, they can do a lot of productive work with the machines and themselves. That’s another way of seeing the value of both of the tools.

How it can improve garments manufacture

via

Computer scientist Alexandre Kaspar mentioned cool things. That it can get used to providing customers with proper designs, they wish for. They can be supplied with the tool, make their designs and send to for knitting. This is a more efficient way to make garments. In the meantime, it can lead to reducing waste products development in industries.

He added that it is like working with 3D printers. Here, actually you work with wool and yarn, instead of plastic. He says 3D knitting is the second generation of 3D printers, that we might have thought of using now. Finally, we can design our own clothes and woolens.

LEAVE A REPLY

Please enter your comment!
Please enter your name here