A few years ago, the term “artificial intelligence” or AI was reserved mainly for science fiction movies. Today, AI is all around us. From self-driving cars to social media monitoring to that virtual assistant you just had a conversation with on a retailer’s website, AI is quickly becoming a part of our everyday lives.
The AI software market is predicted to hit $62 billion in 2022. The largest increase is projected in the area of knowledge management, also known as machine learning. This is the area of AI where machines can use the data they collect to perform tasks and processes.
One area artificial intelligence is seeing growth in is the textile industry. AI is finding a home with textile manufacturers, helping with visual inspection jobs like color matching and pattern making. And some companies are using artificial intelligence to assist with quality control, supply chain management, and an overall improved customer experience.
A new special issue of the AATCC Journal of Research (AJOR) shares papers from the
2nd Artificial Intelligence on Fashion and Textile International Conference.
One area of the textile manufacturing process seeing an increased use of AI is in pattern making. Jeffrey Joines, department head of textile engineering, chemistry, and science at North Carolina State University Wilson College of Textiles, says AI and machine learning can be used within textile machinery to make sure pieces of apparel or other goods are cut correctly so little to no fabric is wasted.
AI can also help make sure patterned fabric pieces are cut correctly so designs like stripes and flowers are exactly where a manufacturer needs them.
“If you can optimize not only how to lay them out to reduce waste, but also to minimize the amount of time it takes to cut on the path they take, that’s also important,” he explains.
And as fabric can sometimes stretch out during the cutting process, Joines says artificial intelligence can also aid with laying out the pattern for fabric cutting. “The vision system can look at it and adjust the pattern just a little bit depending on the actual pattern for that particular piece of fabric,” he adds.
Another area where artificial intelligence is being used in the textile industry is in color matching. According to Ken Butts, global key account team manager for color management workflow company Datacolor, the company’s SmartMatch uses AI and machine learning to automate the dye formulation process.
Traditionally, formulating a dye recipe to match a specific color would be done visually and would usually need a few color correction stages along the way. With SmartMatch, Butts says, the software stores and uses past experiences to produce a color match with a lower Delta E CMC, helping to minimize color correction steps.
“When you put dyes together, unexpected things generally happen,” Butts explains. “There are interactions between the dyes that only occur when they are in combination with one another. Or maybe we’re using a particular material or a slightly different process that causes the color to be produced a little bit differently. SmartMatch basically searches through the database to find any similar colors on the same material, and then it gives us a prediction based upon how those dyes actually perform together. The end result is that our first formula is almost like running a formulation and two corrections all at once because it’s learned about how those dyes work together.”
One area Joines sees growth for AI in the textile industry is in quality control. For example, over the last 10 to 15 years, he says, more machinery manufacturers have been placing sensors in machines like yarn making, weaving, and knitting machines to monitor production issues. In the past, those sensors would stop a machine for a human to come in and fix the problem. With newer AI technologies, Joines says those same machines will use sensors to observe the manufacturing process to make sure there are no snags in production. For example, a machine could “learn” to adjust draw speeds to make sure there are no yarn defects during yarn manufacturing.
“You’re going to see more of these sorts of technologies used, especially in yarns, where humans don’t have to interact to fix problems,” Joines says. “They’re predicted before they happen based on all the sensors that are on the machines.”
AI also assists with quality control when it comes to color matching. Butts says AI tolerance, in the quality control program Datacolor Tools, can be trained by customers who need to have tight control over their color matching to flag production batches as pass or fail based on their specifications.
“The program has been used very effectively by companies who are manufacturing materials that are replenished throughout the year, but also for companies that are multi-sourcing,” Butts adds. “If you’ve got a multi-source program, or if you’re producing components that need to match each other, they’re effectively using the AI tolerance to get much better color control and consistency.”
As AI continues to play an increasing role in the textile industry, Joines believes manufacturers will start using the data their machinery sensors collect internally rather than only making it available to customers. “You’re going to see them integrating that into their own control boxes, monitoring the data to make better decisions,” he explains. “These algorithms are getting fast enough, and the data is coming in, where I think you’ll be able to make changes on the fly automatically, without human interaction. I think you’re going to see a lot more of that.”
And Joines also says AI will allow textile manufacturers to create a “digital twin” of their entire supply chain to help automate and manage the supply chain more efficiently. For example, a yarn producer who needs to send fabrics can watch what’s affecting the apparel manufacturer who contracted those fabric pieces in case there’s a delay somewhere in the supply chain — such as a hurricane — and what steps need to be taken to resolve any issues, such as rerouting the fabric to a different port.
“You’re going to see, in the supply chain area, more use of this sort of digital twin — of being able to automate and manage the supply chain a lot better and more efficiently,” Joines adds. “It’s really about how do you reduce cost — it really comes down to waste and cost. It’s better for the economy (and) better for the environment.”
Author: Corrie Pelc
This special issue of the AATCC Journal of Research (AJOR) shares papers from the 2nd Artificial Intelligence on Fashion and Textile International Conference 2019 (AIFT 2019), which was organized by the Institute of Textiles and Clothing, The Hong Kong Polytechnic University and The 19th International Exhibition on Textile Industry (ShanghaiTex 2019) in Shanghai, China, November 25-27, 2019.
Four key artificial intelligence (AI) topics were discussed at AIFT2019:
Wai Keung (“Calvin”) Wong, Cheng Yik Hung Professor in Fashion, Institute of Textiles and Clothing, The Hong Kong Polytechnic University and AiDLAB Centre Director, was the guest editor for this special issue.
The AATCC Journal of Research Artificial Intelligence special issue can be accessed by all interested in the topics at no cost at https://journals.sagepub.com/home/aat.