Still OCR: Combine captured text

[Shun]

Hi Mike,

I'd like to OCR multiple sequential pages from a book using the iPhone's camera. I've read in the manual that by turning on the "Combine captured text" option in the Block Recognizer OCR Settings, all captured text will be appended to the previous text in a session. Unfortunately, more often than not, the last recognized text replaces all the previously recognized texts, forcing me to start over. I took care to re-enter the Still OCR screen right after the capture, so the OCR session doesn't start anew.

It's great to have the function of appending to the previous text, but perhaps it needs to be supported by the user interface a bit more, for example by asking the user whether to start a new Still OCR session or to append the text to the previous one? Or maybe by offering the user a way to take the next photo from the screen displaying the captured text/after going back to the Block Recognizer screen? Perhaps you have an even better idea.

Another thing, if I turn on the flash in the "Take Picture" screen, it forgets that setting for the next photo. For OCR'ing longer texts, could one make that selection stay within a session? Admittedly, it can be annoying to the user if an app remembers a setting he/she doesn't want anymore, but within a session, that shouldn't be a problem.

Happy New Year, Shun

 

[mikelove]

"Combine captured text" only works within a single image, and only when you tap on the reader button - is that what you're using? Capture some text, tap reader button, go back, capture some more text, tap reader button, go back, etc. It might be better if we had an option to aggregate text over a longer period of time but that's all that that option does at the moment - the intent was that you could capture a few separate paragraphs of text even if the OCR system refused to recognize them in one go and then save them all to one text file.

I didn't think "Take picture" flash settings could be persisted but on checking the documentation it seems like they actually can be, so we'll see what we can do on that for our next update.

 

[Shun]

Ah, so I kind of fooled the system by taking another picture in between. Thanks for the clarification. I would certainly welcome text aggregation over a longer period of time, but I will just number them and concatenate them for now.

Great, thanks.