I'm curious how well the iPhone OCR actually works. How do you deal with errors? Is the error rate low enough that you can accept the output from the iPhone OCR as is or do you also run it through a cleaning process (e.g. spell check)?
For a single data point, it works exceptionally well for me. I routinely copy-paste from images or screenshots, and it rarely fails (mostly for handwriting or obscure fonts).
I am not sure if the Photos app search also uses the same OCR. But sometimes I search for a word and iOS will find a photo that has that text in a 50x20 cluster of pixels somewhere far in the background...it's remarkably good.
