If we would stop producing clothes now we would have enough to last the coming 10 years and everyone could wear something different every day.
So what if machines could learn a sense of fashion and use this to distribute second-hand clothes using the internet?
Fashion is the second most polluting industry in the world (after oil) and faces many sustainability challenges: ‘fast fashion’, ‘deadstock’, ‘sweatshops’ and ‘over irrigation’ are a few key examples. To make it worse, every year each person disposes 24 out of 40 items with non-textile items and these get incinerated. (Measuring the Dutch Clothing Mountain) Out of the other 16 that are separately collected, 11 are suited for reuse.
We are inspired by the way that some shops/stylists are able to pick out, re-market and thereby sell used clothing again. One could argue that so much fashion has already been made and, as fashion trends are cyclical, it’s a matter of knowing how to dig for the right clothes and recommend them to the right person. Our vision is to get maximum use out of all clothes and we are therefore working on algorithms that can recognize, categorize, match and recommend them to people based on their style.
Another aspect is the digitization/photographing of the clothes. We are developing an installation that allows traders of used clothes to do this at marginal cost per item and at webshop quality.

Our first app is now available for testing with the inventory of one of our partner stores in Rotterdam. This app let’s you choose the colors of the clothes and finds you matching tops and bottoms. You decide which ones you like and create training data for a matching algorithm.

setmatchingtest