The brand name I know is BrainPort, which gives you a 20x20 matrix with three intensity values (on, mid, off) on your tongue:

https://www.wicab.com/brainport-vision-pro

I found a few others while digging to find this again.

I can't say how well the translation of the picture into these matrices works, but expect these devices will also benefit from the research into the phosphene stimulation matrices, which (to the best of my current knowledge) have the same pixel matrix resolutions and thus the same problems to solve. A substantial part of the research aims at extracting meaningful pixel matrices from regular cluttered visual scenes through deep learning.