Jump to content
Nitro Victim

Incredibly bad OCR

Recommended Posts

Nitro Victim

I've struggled with how inaccurate the OCR is for quite some time, but I finally received a document where it mattered so much that I did a comparison, a scanned legal (application) document.  

Nitro spent a bit of time processing the OCR... and couldn't find any cases of "NE 10th", which are all throughout this.  Nor "NE10th".  Nor NELO, nor NEIO... I was trying ever-so-hard to determine what it could be seeing.  And only two cases of "10th" at all, as part of a larger address.  

So I imported the PDF into the free NAPS2 scanning software.  NAPS2 is a really great scanner, but also has an OCR component.  I'd never tried this before, but it can indeed import a PDF.  And then I had it create, from the PDF, a PDF.  At which point it OCR'd it.  And found the NE 10th instances.  (And 10th instances too.)

Dear GoNitro... I paid money for this product.  Why can't it work at OCRing PDFs as a free product does that wasn't even designed for that?

The PDF is at https://development.bellevuewa.gov/UserFiles/Servers/Server_4779004/File/pdf/Land Use/19-105108-LP.pdf

Regards,

   A Long-Suffering Customer (yeah, I was bit by that slowdown bug too.)

 

Share this post


Link to post
Share on other sites
Leslie V.

Hello @Nitro Victim,

Thanks for reaching out to us!

We apologize for the inconvenience. I created a support ticket on your behalf so we can assist you directly.

Cheers!

Share this post


Link to post
Share on other sites
CyberRon
On 2/28/2019 at 8:35 PM, Nitro Victim said:

I've struggled with how inaccurate the OCR is for quite some time, but I finally received a document where it mattered so much that I did a comparison, a scanned legal (application) document.  

,,,,

Dear GoNitro... I paid money for this product.  Why can't it work at OCRing PDFs as a free product does that wasn't even designed for that?

2

I am researching Nitro Pro version 12 for a possible upgrade from ver 7. The only real improvement I can see is the creation of a PDF from the clipboard.

However, the improvement is offset by a terrible OCR capability.

I printed a webpage to PDF using the accompaning PDF printer driver for each version. Then I opened the two documents with their respective version of Pro and ran the OCR function. Version 7 worked like a charm - all words searched for were found - all instances.

As for ver 12, after running OCR once, I could only find one of three instances of one word and none of the other. Then I ran it again and I was still unable to find all three instances of each word.

The speed of the OCR was about the same, but the results were vastly different. I then attempted for open the v12 document in NitroPro 7 and could not find the words there either. When I attempted to use the OCR again, I got an error message that OCR was not able to run - the reason - the pages (all 5) already contained searchable text.

I then opened the ver 7 created and OCR'd document in ver 12 and was able to search and find all instances of the two words I searched for.

Looks like Nitro Pro has taken a few steps backward from Ver 7.5 - at least in the OCR department. Since I use PDF's for archiving important documents, OCR is a vital element. It looks like I will have to look for another program if I want to upgrade my Ver 7.5 Nitro Pro for some of the other features that I crave. I still have not found anything that does as well as Acrobat, but for personal use it is out of reach - for now.

I can upload both documents if it would help...

Edited by CyberRon
Offer for uploading documents in question

Share this post


Link to post
Share on other sites
CyberRon

I tried the Searchable and editable feature (after trying the searchable only option). I was able to find two words I was looking for, but though they each only show three times in the document, the find function found 6 instances of the first word (three on the second). They seemed to show one instance of all the letters except the last one and then all the letters of the same word in the second find for each instance.

Further, by using this S&E option, the text in the document becomes fuzzy - not as sharp as before the OCR. Not very useful.

This is in the latest trial version of Nitro Pro downloaded yesterday - version 12.10.1.487. Also, check my other posts about the trouble I am having with OCR on this version compared to version 7.5.0.29. I was hoping to upgrade my version, but I am not paying for a downgrade!!!

Share this post


Link to post
Share on other sites
Steven Bailey

I really don't understand how perfectly clear documents, which consist only of Bookman or Times Roman text for example, with no formatting other than paragraphs, can get messed up so badly by clicking on "make this document searchable". I didn't realise that so many of my documents were altered until after I'd deleted the ones which I'd "combined" into pdf.  lo changed to b for example and other changes making documents unreadable.  This "in product notification" should be turned off by default, and a dire warning dialog added for the benefit of those that may be unaware that their text is about to be destroyed.
 

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.