Time Chain, (2020)
How long remain traces of our movement in time? What does it look like? Digital images, art prints and GIF loops based on machine vision techniques tries to answer these questions and raises more. Traces of people, cars, birds and things emerge from the recorded images and reveal uncanny semi-random structures.
I used the RCNN neural network and OpenCV to mask selected objects out of the video (or image sequence). These masked objects (like people, cars, etc.) were separated from the background and further manipulated, so they create “time chains” - traces of objects in time. Each object is also converted to black and white dither, with many random modifications determined by time and type of object. Dithering enabled me to create relief prints using laser cutter and traditional woodblock printing.
Artworks are reflecting my own personal experience with collecting and viewing data using machine learning techniques. Our movements in public space are constantly monitored and evaluated by similar machine vision algorithms. What if we try to use these tools to trace birds in flight, snow falling and other natural movements? My intent is to examine this topic more from an artistic point of view and use the data and algorithms to generate unique high resolution “digital drawings”.
Machine vision is inevitably connected with glitches and noise (which I intentionally work with). Incredible amounts of data leads to information overload and inability to reasonably use it. I would like people to ask questions like: Should we blindly trust these surveillance algorithms? What happens inside the machine, how do “they” see “us”? There is also certain amusement of simply seeing things, which we can not directly observe in real life.Back To The Gallery