There is a neuromorphic vision sensor that is generating some interesting research these days that is called the Dynamic Vision Sensor. One of the claims of this sensor is good approximation of the conventional retina. However, there's no detail on the website in what way this is true or to what degree this is true.

It does not detect colour, so obviously it isn't an accurate approximation of the cones found in the retina. Additionally, the function of the cornea is fulfilled by the lens of the camera, so this input isn't totally approximated. Consequently, the only significant component that remains to be evaluated is the rods of the retina?

What are the similarities and differences between a retina and the DVS sensor?

First, let me save you some reading by explaining how the DVS sensor is different from a typical camera. Instead of synchronously sending frames with data about every pixel, the DVS is event driven. That means, when a light intensity change occurs, an event is fired with the address of that change, a time-stamp and whether the pixel observed an increase or a decrease.

Do rods also encode changes in light intensity in the same manner? If not, what is the difference?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.