Facedeep
August 2022 - February 2023
Facedeep is an ongoing investigation into facial expression - and how to interactively visualize algorithms. My algorithm uses the python package "deepface" to live analyze the emotions of people interacting with the installation.
"Deepface" assigns values to 7 different expressions ("angry", "disgust", "fear", "happy", "sad", "surprise" and "neutral") These 7 values are then used to create one line of the graph. Everytime a new expression is captured, another line is added to the graph - allowing the observed to draw through their facial expressions.
The expressions however are only facedeep - they don't relate to the actual emotional state of the observed. Unlike most humans, the algorithm doesn't understand the potential situational cues, it cannot see what it hasn't been given as an input. All it can do is draw conclusions about predominant emotions based on what facial expression it sees - the visualisation is supposed to help the observer understand that the algorithm is far from all knowing, and to understand the connection between input and output.
The expressions however are only facedeep - they don't relate to the actual emotional state of the observed. Unlike most humans, the algorithm doesn't understand the potential situational cues, it cannot see what it hasn't been given as an input. All it can do is draw conclusions about predominant emotions based on what facial expression it sees - the visualisation is supposed to help the observer understand that the algorithm is far from all knowing, and to understand the connection between input and output.
Year 3 - Fundamentals
XY Plotter, Python (deepface, vsketch, axidraw), Literary Research