• Brooke Roberts-Islam

How fashion graduate Mathilde Rougier is using AI AND AR to eliminate fashion textile waste

Updated: Feb 14

Mathilde Rougier is a recent graduate of Central Saint Martins, where she studied on the womenswear pathway. For her final collection, she wanted to create new designs from old garments and samples of fabric she had acquired from past internships. This circular approach to fashion – where textile resources remain in repeated use rather than going to a landfill or being incinerated – is one that the industry as a whole must move towards, considering the amount of textile waste it produces. Like Mathilde, many students from Techstyler’s Digi Fashion Class of 2020 showcase have incorporated sustainable/waste-reducing practices into their collection, showing an acute awareness of the climate emergency and the incredibly wasteful and pollutive industry they will inherit. It’s a dilemma many fashion students have tried to come to terms with: loving fashion creation but also knowing that the world does not need more clothes. To overcome this problem, Mathilde decided to adopt innovative techniques.

She took photos of the old garments she wanted to use, and these became the basis of the augmented designs in her final collection, which is a tessellation of leftover textiles and existing garments, re-imagined virtually. Although she was interested in 3D design before and planned to use it in her final collection to some extent, Mathilde is one of the many fashion design graduates for whom digital fashion took on a much larger role in their projects due to COVID-19.

It must have been bittersweet to have your last year cut short; how did you feel? At the beginning, when lockdown was announced, I had a couple weeks of uncertainty. I didn’t know what to expect or how long lockdown would actually be. So for that period I was a bit of an emotional mess. Then I became very determined to make the situation better for UAL students: with petitions, meetings with university staff, demanding to cut the cost of tuition… This eventually fizzled out because we were getting nowhere. Finally, after that, I began to feel more acceptance. I came to realise that I have the opportunity to really throw myself into something which I had wanted to explore before. I was in a good position as the main equipment I needed was a computer. I already had the squares of second-hand materials (from sample books I had collected over internships to avoid textile waste) I was going to use to map the pixels of the digital garment. The only machinery I needed was a heat press for the recycled plastic I was using.

How did you overcome the obstacles of designing in lockdown? Was it difficult to stay motivated and in touch with your creativity?

Actually, I thought it was quite exciting in a sense. I was more productive. It may have been the sense that it was all in my hands, no-one was going to push me. The only real problem at the beginning of lockdown was shipping all my stuff back to France, since I was going to spend lockdown with my family. I needed to ship all my squares, all my materials. I couldn’t replace them because the goal was for them to be recycled materials in order to minimise waste.

What led you to choose digital fashion design over traditional methods? Was it an unexpected solution to the limitations of COVID-19 or had you been interested in 3D design before?

I was interested in it before, but it took on a larger role because of the physical limitations of COVID. Digital fashion had already been part of the plan, but not to that extent. What I do is augmented reality, so a middle ground between physical and entirely digital fashion (with 3D avatars). I still want to dress people but want the option to do so without producing waste. You can update your designs constantly without waste.

What was the inspiration behind your collection?

It didn’t actually come from an aesthetic inspiration, more a problem-solving approach. My approach was technical: I wanted to address the issue of sustainability. I used different tech systems (whether that’s pixelation*, convolutional neural network**, etc.) to solve problems to do with sustainability.

* The images of the old garments were pixelated in photoshop or sections of the garment were 3D scanned. This allowed Mathilde to translate the real-life textiles into a virtual medium for use as the basis of new 3D designs (i.e. the old garments would be the foundation for the Augmented Reality digital layer of Mathilde’s designs.)

** A convolutional neural network is a type of deep neural network that analyses images to determine and categorise their visual characteristics, in effect recognising details of the garment – such as edges or hard components like buttons – and allowing the designer to adjust the pixel patterns within these garment details. The significance of being able to adjust the pixel patterns is two-fold: first, ensuring that everything looks as intended and to make the transition from physical to digital more seamless, and second, it gives the designer the option to infinitely rearrange and “play with” the building blocks of their design.

What software did you use in creating your designs? Were there any challenges?

I used a convolutional neural network (AI), Spark AR and Blender (3D animation software). I learnt how to 3D model during the project, so it took some time before I was comfortable. It was a challenge to try and get the AR right and plane-track* the clothes. The thing with 3D design is that you need to make sure the design looks good from different angles. This means blocking out the body and using occluders** to make sure that patterns on the back didn’t show up from the front and vice versa. It took a lot of adjustment to make sure it worked.

*Mathilde used plane-tracking on second hand garments to add an Augmented Reality digital layer on top, using software such as Spark AR and Blender. Plane-tracking involves taking a photo of a recognisable high contrast area of the garment, which is the bit that acts as a trigger that causes Augmented Reality interaction.

** Occluders are objects which impede the amount of light that reaches the eye. In this context, an occluder would block a pattern on one side of the garment from bleeding into the other side. Mathilde had to put explicit instructions into the software to ensure that everything was visualised in the correct location and orientation, according to Mathilde’s design.

This is an example of what could happen if occluders are not used: you can see the details of one side of the 3D object bleeding into the other.

What is your plan after graduation? How do you feel about going into the industry during this uncertain time?

I don’t really know yet. At the moment I’ve had lots of freelance jobs which I don’t think I would have gotten 6 months ago. All of a sudden people are interested in digital fashion, there’s a lot of hype. COVID accelerated things because everyone was in front of their screens. But people have been experiencing fashion digitally for a long time without really realising it: e.g. all the people that watch shows online rather than being there in real life. Fashion enthusiasts have been consuming fashion from their screens for ages, but now, since everyone was in lockdown, industry professionals and the fashion system have had to adapt a bit more. I’m going onto MA Accessories at the Institut Français de la Mode in Paris so I’ll carry on freelancing on the side, but I’m lucky to not have to think about a full-time career yet.

How do you want to see the fashion industry change? Is it already changing or is this evolution merely superficial (for example in terms of sustainability)?

Fashion needs to move towards a circular model of production, i.e. stopping with the production of new textiles and making use of existing garments and materials. In terms of sustainability I think we’re still quite far behind. Sustainability is such a large word, there are so many ways to go about making the industry more sustainable. In a sense it’s a good thing as there are many small ways to help, but the risk is focussing too much on the small things rather than the big problem at hand. We see a lot of greenwashing and superficial fixes to a problem that is fundamentally wrong. I’m curious to see more systemic change.

Is 3D fashion design the future?

It’s quite revolutionary waste-wise for smaller brands since the proportion of sampling to product output is bigger than for large companies, meaning it would be extremely worthwhile to move to virtual sampling. With this, you reduce waste at the design and sampling stage and can make many alterations on a toile.

What about digital fashion from end-to-end (meaning no physical product is made)?

This would work for brands doing digital showcases. I have a preference for augmented reality (meaning real people wearing digital clothes rather than a 3D avatar). I still want to dress real people, but people have increasingly developed a digital aspect to their personality. For example, influencers on Instagram that wear a whole new outfit for every new post: augmented reality could solve the overconsumption issue caused by this. The medium is digital so why not the clothes? It can be integrated into the way we live our lives already.

Lastly, what do you think digital fashion means for sustainability?

It allows design without physical production, therefore less waste. Though you can’t totally ignore the carbon footprint of technology, the electrical energy used in powering a computer would be less than trying to recycle all the waste material and mixed fibres at recycling plants. Digital fashion at least streamlines the waste and carbon footprint caused by the industry. Also, there’s visibly “tech-y” fashion (virtual fashion) which will certainly affect the zeitgeist and our opinions on owning physical clothes, but I think the core change in terms of sustainability will be about incorporating technologies into making physical products. You might look at a T-shirt and not think there’s anything particularly innovative or futuristic about it, but the truth is it was made using digital processes that aid sustainability.

With the goal of a more sustainable fashion industry in mind, Mathilde wanted to address both our physical and digital identities. This is due to our digital identities playing an increasingly bigger role in our lives, and these digital identities (or rather the social media sites where they live) being ruled by algorithms that constantly call for new content. She remedies this in the physical realm through the fabric squares she used for her designs which can constantly be repositioned, creating new designs without creating more waste. This translates to the digital realm too, where her Augmented Reality virtual outer shells (each of which correspond to a different garment) can continuously be updated. 

Mathilde enjoys being able to design without creating any physical waste and strongly believes the waste issue in fashion is one that needs to be addressed by the industry. However, she doesn’t want this to come at the cost of design, whether that’s aesthetically or in terms of speed. This is what her collection tried to remedy, showing that zero-waste, repurposing and recycling techniques do not have to produce the “crafty” results they are often associated with. Like much of fashion’s “new guard”, Mathilde is an advocate for the circular mindset where things can constantly be re-designed, allowing for full creative expression by the designer and satisfying the fashion crowd’s love for newness without creating more textile waste and contributing to the devastating climate impact that comes along with fashion production and consumption. In this way, the designer can fully explore the best of fashion – unbridled creativity – whilst avoiding its destructive impact.

By Anastasia Vartanian