Tools to analyze and compare images, focused around discovering, understanding and documenting street art.
Enter your email to get pleasantly infrequent updates, on new blog posts and releases.
Modern machine learning tools have made it possible to analyze and manipulate large datasets. This has especially improved how images can be queried and searched.
Using techniques that were previously only available in academia or private industry, Public Art hopes to use machine learning techniques to better understand street art on a large scale. Public datasets in the form of social media and user generated content provide enormous image repositories that make this possible.
Below are six practical aims for applying these techniques.
Publicly labeled street art can be used to identify who painted a photographed mural.
Analyzing images around the world can help track where certain styles emerge and how they are adopted.
Images without any metadata can be correlated to geographic data based on visual landmarks.
Computer vision encoding techniques can extract image features and generate visually objective categories.
The content or subject of street art murals and tags embody social phenomena in an impermanent medium. Digital documentation can provide a retroactive look at these historical moments.
Monitoring newly posted images online can help detect when a new painting has been photographed in an area.
Want to know more about street art and machine learning? Enter your email to receive pleasantly infrequent updates.