Capturing the expressions that most confuse machine learning classifiers. Mapping beyond the bounds of classification allows users to create an intimate understanding of their own spaces of expression.

Join us in a distributed collective action to reclaim emotional nuance through unconventional interaction with facial recognition AI. The datafication of every single action or behavior online has led to the progressive institutionalization of all aspects of life, including our emotions. Facial recognition software is designed to identify, sort and classify based on a particular set of predefined moods. By turning these limitations into material for creative reuse, our goal is to generate intimate emotional portraits composed of expressions that cannot be easily classified in that they belong to a much richer spectrum of human experience. Using webcams, the Emotional Bounds project consensually observes participants over time while they work at their devices and records only unclassified, uncertain, and undefined facial expressions of the photographic subject. Images are not collected or stored but rather persist only if the participant chooses to download their data set and recontribute it to the project.

Downloading models...