Does Vision API need preprocessing? Multiple QRCodes

Question

I'm using mobile vision to detect multiple (they're usually 4+) QRCodes in images.

Will pre-processing make detection better? (binarizing the image, anything else?)

In my previous app runs I've almost never got all QRCodes in the frame, any suggestions?

Answer 1

The Barcode API prioritizes speed over reading every single QR code that is visible, so about 3 simultaneous QR results is the rough upper limit for it will return for a single frame.

We already perform binarization, so that will not improve performance. If you have an idea of how the codes will be distributed spatially, you could try slicing the image into tiles.