Most AI vision models are 'blind' to optical illusions that fool every human, revealing a massive gap in how they process motion.
April 14, 2026
Original Paper
Do vision models perceive illusory motion in static images like humans?
arXiv · 2604.09853
The Takeaway
State-of-the-art optical flow models fail to see movement in static images like the 'Rotating Snakes' illusion. Only by implementing human-like dual-channel processing and saccades can models reproduce this effect, proving AI vision lacks a core biological temporal component.
From the abstract
Understanding human motion processing is essential for building reliable, human-centered computer vision systems. Although deep neural networks (DNNs) achieve strong performance in optical flow estimation, they remain less robust than humans and rely on fundamentally different computational strategies. Visual motion illusions provide a powerful probe into these mechanisms, revealing how human and machine vision align or diverge. While recent DNN-based motion models can reproduce dynamic illusion