People You May Know
Abstract photographies composed by blending images from Facebook friend suggestion albums into single pictures. Using custom made software, each original image selected by the Facebook algorithms is stacked on top of each other, and different sets for different human body features as detected by feature detection algorithms are condensed into separate emotional impressions by processing their mean color value per pixel. Thus, each picture is a digital mixture of both algorithms as they work around our digital presence.
Each final image captures the logic of human body feature detection algorithms but also fuses it with a distant, shadow view of digital body feature representations of people "You May Know", selected by another set of Facebook algorithms running on the shadow-profile user interactions. All those uncanny recommendations are digital shadows that constantly follow our virtual avatars, as they move around inside cyberspace, casted by the constant light of network activity surveillance.
The digital images were created by bulk downloading 9000 images downloaded from Facebook friend suggestion albums for my personal account. All images were then stacked using custom made software to process their mean color value per pixel, and increase the dynamic range of colors retaining the highest of their original detailed variety. They were also centered and rescaled at the same size of different human body features, as recognized by feature detection trained neural networks using the OpenCV library.
Using a high number of original images the effect of the body feature detection algorithm is much more pronounced, but by using a lower number of images it is possible to make the Facebook friend suggestion algorithm effects more visible. The final image captures different symmetries between those two algorithms, depending on the number of images used each time.
Fullbody.
Fist.
Hand.
Lowerbody.
Mouth.
Upperbody.
Exhibited on the featured artists section of The Wrong - New Digital Art Biennale 3.0: "Anestis Anestis" (November 2017 -January 2018)
The digital images were created by bulk downloading 9000 images downloaded from Facebook friend suggestion albums ( http://www.facebook.com ) for my personal account. All images were then stacked using custom made software to process their mean color value per pixel, and increase the dynamic range of colors retaining the highest of their original detailed variety. They were also centered and rescaled at the same size of different human body parts, as recognized by feature detection trained neural networks using the OpenCV library.
Software written in Processing. Processing is an open source programming language and environment, originally built in the Aesthetics and Computation Group at the MIT Media Lab as a domain-specific extension to Java, targeted towards artists and designers. http://www.processing.org/ http://www.processing.org/exhibition/
Feature detection programming based on the OpenCV library and pre-trained Haar feature-based cascade classifiers. OpenCV (Open Source Computer Vision Library) is an open source computer vision and machine learning software library. http://opencv.org/












