Scientists Crack How Your Eyes Detect Speed and Direction
The same math could soon help self-driving cars and phone cameras see motion.
A researcher named Tony Lindeberg has published a theoretical analysis of how idealized models of the spatio-temporal receptive fields of simple cells and complex cells respond to motion, based on the generalized Gaussian derivative model for visual receptive fields. In the model, receptive fields are represented as velocity-adapted affine Gaussian derivatives for different image velocities and degrees of elongation, and they are probed with moving sine waves at different angular frequencies and image velocities — a method structurally similar to how biological neurons are characterized. Comparing the theoretical results with neurophysiological measurements of direction and speed selectivity for biological neurons in the primary visual cortex, the paper reports consistency with velocity-tuned visual neurons that are sensitive to particular motion directions and speeds, and with different visual neurons having broader versus sharper direction and speed selectivity. The results, combined with neurophysiological characterizations, are also consistent with a previously formulated hypothesis that simple cells in the primary visual cortex ought to be covariant under local Galilean transformations, so as to enable processing of visual stimuli with different motion directions and speeds.
- It explains why some brain cells fire only for one direction and speed of motion, while others respond broadly — useful for spotting both slow buses and fast balls
- The math matched real recordings from animal brains, giving scientists a working model of motion vision
- The same equations could make self-driving cars, video streaming, and phone cameras faster and more power-efficient, though nothing is shipping yet
Why It Matters
Could lead to cheaper, sharper motion vision in cars, cameras, and robots — and new insight into motion blindness after stroke.