Andrés Caballero

Tracing Lines
Tracing Lines is a 30x40” photo print that uses computer vision to detect law
enforcement officers from online videos of migration protests. When the piece is viewed
through a mobile device, the image comes to life revealing a short augmented reality video. 

Andrés will work with online found footage of recent protests, where the government has shown a
disproportionate use of force against peaceful demonstrators. Through AI-powered facial
recognition, government agencies identify protesters and track them to their homes for
questioning and intimidation. This piece responds to the way cameras and machine learning are
currently used to sustain a constant state of oppression. Using Python and coding with the help
of AI assistants, I will build software that can distinguish civilians from law enforcement officers
in web videos sourced from YouTube, social media, and traditional news outlets. Once they are
detected, another script will perform a “line tracing”, where it strips away the background and
leaves only thin, bright outlines of officers and vehicles. These lines evoke a dystopian image
that foregrounds the current sociopolitical landscape of the US/Mexico border. The final piece
will be a composite of these outlines, condensing different frames into a large grid. Using a
mobile app, viewers can point their phone at the print and watch the figures come to life,
further emphasizing how technology mediates ‘seeing’. The work asks what happens when we
turn the same technologies that are used to control us back onto those in power. Revealing an
underlying infrastructure of violence and its instrumentalization through technology. It
repurposes the use of machine learning and computer vision tools, offering a critical approach
that reveals their lack of neutrality.

This project relates directly to Andrés' current artistic research examining the
infrastructures that target political dissent and further marginalize communities in the
borderlands. It will deepen Andrés' skills in AI-assisted coding and computer vision. Additionally, it
gives Andrés a concrete case study he can bring into future workshops and classes, showing students
how these tools can be repurposed under a critical approach, understanding how they work
and how they can be redirected toward accountability and resistance. The project is a chance 
to share open source tools, workflows, and resources so others can adopt these methodologies in
their own practices.