However, the complex reflections of light in the space made accurate tracking difficult, and the system did not fully function as intended.
(White noise is intentionally included.)
This project started by installing the fabric and experimenting with different light sources, looking for a balance where the projection remained visible while still bright enough for the subject to be accurately tracked. Ultimately, I had to accept that the reflective light in the space, it was impossible to reliably capture the moving subject.
One aspect that became less prominent in the installation due to technological difficulty was the torch element. The wrist‑tracking data was inconsistent, so I initially added a secondary switch to activate or deactivate tracking based on hand detection. This worked well during early testing, but at full scale the hand was inevitably always in frame, producing messy data. My intention was for the torch to change size along the z‑axis, simulating the experience of holding it in real life, but this further disrupted the tracking and ended up complicating the interaction.
The eye‑tracking ended up working surprisingly well. I masked a shape onto an eye that inverted with each blink. The on‑screen eye mirrored the observer’s blinking speed, reintroducing the theme of surveillance and perception.