Aesthetics Of Exclusion

Aesthetics of Exclusion is a design research project that questions how we can use computer vision techniques and machine learning to explore and analyse aesthetical styles that correlate with gentrification in large image archives such as Google StreetView and Instagram.

In 2007 Google StreetView started recording city life. Each year the appearance of whole cities is documented. Yet few have used this growing archive for research purposes, since it’s not easy to analyse tens of thousands of images. At the same time, the development of computer vision techniques, such as image recognition gets more efficient and precise every year. We can use these techniques to tell the difference between cats and dogs, recognize particular artistic styles, search large historical image archives, and in combination with machine learning, even use it to predict behaviour.

Gentrification is an urban phenomenon and it is reflected in the removal of communities of diverse classes, ethnicities, races, sexualities, languages, and points of view from central city neighborhoods, and their replacement by more homogeneous groups. In short, it is a global process of homogenization expressedon a local level.

While a lot of research has been focused on the socio-economical causes and effects of this important urban process, its aesthetics have remained understudied. Yet, it is exactly the aesthetic signposts of gentrification that can tell us much more about how neighbourhoods are actually changing, and in which stages of gentrification they currently are.

The research process itself will open up a lot of questions and room for experiments. How do we as humans teach a computer to recognize the complicated concept of gentrification?

Can we study how computer vision algorithms make decisions? For example, are we able to isolate the specific objects, colours, or characteristics seen as gentrified? Is there a global style of gentrification or are there significant differences between cities like Seoul, New York, and Amsterdam? And can we use generative Computer Vision techniques or style transfers to predict how streets will look after they have been gentrified?