Random noise with action, creating an interactive illusion!
[Preprint Paper ] [Code ]
You can try the illusion demo yourself with the instructions provided below. Please note that the actual demo is located at the bottom of the page. The pictures and videos in the middle section are instructional videos, not demos. They are meant to guide you on how to experience the illusion effectively.
The quality of the videos below may be limited by your internet connection. However, our data shows that these illusions are robust across a wide range of parameters.
Action: Drawing "S", circles, or spirals on the display.
Observation: A visible trace appears and fades within 0.5-1.0 seconds, seen as either brighter or dimmer.
Action: Opening and closing fingers while moving toward and away from the display.
Observation: Dots seem to move away when approaching and are attracted when receding, with effects intensifying on repetition.
Action: Moving an open palm slowly in small ranges close to the display.
Observation: Dots appear to move with the palm.
Action: Moving hand up and down behind the display.
Observation: The noise seems to respond to the hand's movements, intensifying with focused attention.
Summary:
Interactive Perception: Observers perceive interactive illusions as if the random noise responds to their movements, showcasing an interplay between human perception and visual elements.
Mechanisms Engaged: These illusions involve multiple mechanisms working together, potentially explaining significant differences across participants.
Robustness in Perception: Perception values vary with cell size and temporal frequency, but the overall perception remains robust.
Try different actions with these noise interactions!
Note that this noise is generated by HTML and CSS, so it may not perform well at high frame rates.
Underlying mechanism? - Attention
For the following noise video, use YouTube videos at 1080p 60fps and ensure a good internet connection.
Acknowledgements
The project was supported by SONY Research Award Program. We thank Jiahe Yue for her suggestion to employ a static noise.
Correspondence to sshimojo@caltech.edu and szheng2@caltech.edu.