A new algorithm maps the origin of modes and their propagation in the world

Among clothing designers and brands, the general consensus has always been that the major global fashion centers were New York, London, Paris and Milan. Every year, the greatest designers in the world displayed their products in these places and the colors, patterns and styles spread around the world from there.

But in recent years, other centers have sprung up in Sydney, Moscow, Rio de Janeiro, Jakarta and many other places. The fashion world has become multipolar. And that raises the interesting question of how these places became influential and how fashions are now spreading around the world.

Today, we are getting an answer thanks to the work of Ziad Al-Halah at the University of Texas at Austin and Kristen Grauman at Facebook AI Research. This pair taught a machine vision algorithm to recognize new modes and then use it to search for those styles in a global database of geotagged images.

This process is extremely revealing. It shows when each style first appeared and how it spread around the world. The results reveal a previously unknown complexity in how different cities and countries influence the way fashion evolves, and even predicts how styles will spread in the future.

An eye for style

Researchers start by training a machine vision algorithm to recognize the style of clothing in images. The machine learns to recognize color, pattern, type of clothing, etc. for an outfit, and learns 50 different fashion styles.

Then, Al-Halah and Grauman unveiled this algorithm on a large database of geolocalized images published on Instagram and Flickr. This dataset contains 7.7 million images showing people from 44 cities around the world between July 2013 and May 2016. The algorithm then tracks when and where each style first emerged and how it spread. from there to the world.

The results reveal a complex pattern of influence between cities. Al-Halah and Grauman calculate the degree of influence from one city to another and plot the results on diagrams of agreements which record these connections. Our results highlight the spatio-temporal migration of fashion trends around the world, revealing which cities exert and receive more influence over the others, which most affect global trends, which contribute to the predominance of a given style, and how the degree of influence has itself changed over time, they say.

On the map

The process reveals a number of significant patterns. For example, researchers say that a few fashion poles, such as Paris and Berlin, exert an influence on several cities while being influenced by only one or two other cities. Some cities, like London and Seattle, act as focal points for fashion styles by receiving and influencing simultaneously. And they add that other cities, like Seoul and Osaka, have a high net influence, which suggests that they have unique fashion styles that are not influenced by other cities.

Al-Halah and Grauman continue to reveal that certain cities seem to act in concert by influencing the same set of styles. For example, Chicago, Vancouver and Toronto are a team because they seem to influence similar sets of visual styles.

The influence of cities also changes over time. While some cities like London and Rio maintain a stable influence score over time, others like Austin and Johannesburg show a positive trend and gain influence, say the researchers.

It’s an interesting job. Our findings suggest how computer vision can help democratize our understanding of the influence of fashion, sometimes challenging current perceptions of the regions of the world that drive fashion, conclude Al-Halah and Grauman.

Ultimately, these researchers study how ideas and their physical embodiment propagate in human society. It is an approach that could extend beyond fashion to product design, modern art and even science in general, provided that good databases are available to allow this kind of job. There are many more where it comes from!

Ref: From Paris to Berlin: discovering the influences of fashion style around the world
. arxiv.org/abs/2004.01316