I used ABBYY's Fine scanner to scan my document. Now what?

Hello all,

I will paste here the feedback email I sent to ABBYY concerning me not being able to retrieve OCR text and about FineScanner's accessibility.

I am using VoiceOver and I am not able to retrieve the OCR result from a scanned document both in Book and Document mode. First, when I press the OCR button, check the settings and press Done, I get no feedback about whether and when the recognition is already done. The only way to check the progress is to press the OCR button again repeatedly until the alert that text is still being recognized disappears. However, when it does, instead of showing the resulting text, the OCR settings screen is shown again and tapping Done just brings me back to the scanned image screen. Therefore, I am not able to retrieve the recognized text, even though the same steps worked yesterday, as after several tries to OCR, the text was shown, but not anymore as of now.

Also, I find online OCr rather slow, taking about 1 minute just for a single page. TextGrabber, on the other hand, is much faster with its offline OCR, although it lacks text defects correction features available in FineReader.So, there is no app which would take the best from both of these apps. I would also be grateful if the accessibility of the apps was better, such as proper button labeling and feedback.

Anyway, thank you for your efforts and for making quality OCR software and I wish you the best.

Hello again,

I will also share my small experiment comparing FineScanner Pro, knfbReader and TextGrabber.

I took a photo of an opened book with the standard Camera app, that means without any software guidance tools, such as those in knfbReader, nor with any human help.. Then I loaded the photo to each of these appss, run OCR on the image and compared the results so that the same image is processed in all the apps.

TextGrabber had precision which I would approximate to 35%, although in better conditions when I scanned a flat, clear and well lit document, the precision was almost 100%, specifically, the only error was in some words near the bottom edge of the document where the higher shot angle is starting to affect the readability. FineScanner turned out to have the best results, with precision about 98% compared to about 90% of knfbReader. Though it's worth noting that FineScanner uses online OCr and the processing takes much longer, about 40 seconds in my case but it may differ, compared to about 3 seconds of offline OCR of TextGrabber and knfbReader.

However, FineScanner may be preferred since, as the only one, is able to process more images at a time when Images are selected from the camera roll, though Braigo Companion, another OCR app which is free of charge, can do that too. KnfbReader have a batch mode to capture more photos and process them at once, but cannot do this with pictures from the camera roll. Another advantage of FineScanner is that it is able to also recognize structured and formatted text like headings, lists and tables, and save such rich text to formats like Microsoft Word or PDF.

The bottomline, at least from my experience, which may differ from your conditions, is that if you want the best OCR and don't mind a higher cost, then knfbReader is the right choice and is the one which is the most easy to use and the fastest to retrieve the text which can also be easily read back and navigated within. But if you just want to discern what is a paper or grocery package in your hands about, or what's roughly going on your computer screen when VoiceOver is not reading en error message, then TextGrabber is an sufficient option. Unfortunately, as of now, I cannot recommend FineScanner due to the reasons mentioned in my previous comment, but I encourage you to send feedback to [email protected] or better yet right from the app to kindly ask for resolving this issue.

I greatly appreciate this posting. I was about to buy it and se what happened but now I am just going to do the KNFB reader instead because I am not wasting money on something that just does not work.
Thanks again for doing this!!
A

You are welcome, however bear in mind that this might be a temporary issue. If it is a general problem, then ABBYY are probably currently working to resolve it. So from my experience, FineScanner is not the fastest one, but probably the most precise one.

Have a nice day

Hello,

after contacting the ABBYY's support, we found out why OCR seemed not to be working. The reason is that OCR won't return any result if there are too many defects in the image, even though it would actually be able to recognize at least pieces of text. It looks like there is some threshold of the amount of text recognized which tells when it is worth of returning something. This measure makes FineScanner quite inappropriate for visually impaired users, as they can hardly capture the image properly without defects, and because they often just want to find what is the document roughly about, as even bits of poorly recognized text can help them.

Another thing is that FineScanner doesn't provide any feedback to a VoiceOver user to notify them that OCR has been finished or that not enough text was recognized to retrieve any result. Thus you have to repeatedly check the OCR button at the middle bottom of the screen, which will either show an alert that the document is still being recognized, or open a new screen with the recognized text, or again open the OCR setting screen, which means the server decided not to return anything worth of retrieving.

I was told that my suggestions to let the app retrieve at least pieces of text and to provide better feedback to a VoiceOver user will be taken for a consideration. So, with respect to the fact that FineScanner turned out to have the best OCR precision in my test, we might find this app to be a good choice in the future.

Best regards
Adam Samec