Making full text OCR results accessible outside of search

[@lpasqualis 0]

Hello,

given that Evernote already has a solid OCR engine that examines pictures to make the text searchable, I am wondering why not making the full results of the OCR analysis available as editable text with or along the note itself.

 

I'd like to be able to click/tap on an image and get an option to see and copy/paste the full text extracted from the image, and perhaps to be given the option to append it below the image itself in the note. 

 

This feature would be a fantastic addition IMHO.

 

Is this something that is in the works?

Or perhaps this feature is already available and I simply haven't found a way to access it yet?

 

Thank you.

 

 

[gazumped 11,652]

Hi.  Welcome to the forums.  Your suggestion is interesting,  but I suspect too difficult to implement.  The OCR process for pictures builds a 'tree' of words since the quality is often not good.  House could be horse or hearse.  There are so many possibilities that it's probably not feasible to string the content together as a coherent piece of text.  PDF files also have a 'layer' that can be assigned to text content,  where pictures can only show one page of one image.  There's nowhere in the picture to store the content - the 'tree' information is presumably kept on the server separately.

 

You could always add your own keywords to notes containing pictures as pointers to the content,  but if the OCR process is picking up the wording reasonably accurately,  you should be able to find them directly with a normal search. These posts do get read by the developers,  so if they think its a worthwhile change I'm sure we'll see something like this in due course...

[bduncan 49]

I only have the Mac client in front of me at the moment so I am not sure if Windows supports this, but you can right-click on a PDF and choose Save Searchable PDF As.

It will then save the text content of the PDF.

Here's a screenshot: https://www.evernote.com/shard/s1/sh/abddbb99-b30e-4718-aca6-b38f9796dd61/1c22b8a09738809b60dfc92442ed5b59

[@lpasqualis 0]

Thank you for the welcome and for the thoughtful answer. Very interesting!!

 

If I put my software engineer hat on, I can totally see how that would get complicated given the way - as you described and as I understand - the information is stored (that is, for search and not to get a readable text).

 

That said, if I wear the Evernote end-user hat, I also see how unbelievably powerful that feature would be to me.

 

For example, I often take notes during meetings on my Moleskine and store it in Evernote. The Evernote OCR subsystem does a great job to make my notes searchable (even if I suspect that the text recognition is far from perfect, it is great for search). Often I then have to re-type parts of my notes to include them in documents, meeting notes, etc. If Evernote generated an automatic best-effort editable text that I can then copy and paste, it would go a long way to bridge the gap between paper and electronic notes storage.

 

Even though I understand the difficulty of making this into a feature and my expectations are low, I will still continue hoping that this feature will happen in the future It would change my way of using Evernote in combination with paper notes ;-)

 

Thank you.

[gazumped 11,652]

I only have the Mac client in front of me at the moment so I am not sure if Windows supports this, but you can right-click on a PDF and choose Save Searchable PDF As.

It will then save the text content of the PDF.

Here's a screenshot: https://www.evernote.com/shard/s1/sh/abddbb99-b30e-4718-aca6-b38f9796dd61/1c22b8a09738809b60dfc92442ed5b59

 

Yup.  Same deal in Windows.

[gazumped 11,652]

Thank you for the welcome and for the thoughtful answer. Very interesting!!

 

If I put my software engineer hat on, I can totally see how that would get complicated given the way - as you described and as I understand - the information is stored (that is, for search and not to get a readable text).

 

That said, if I wear the Evernote end-user hat, I also see how unbelievably powerful that feature would be to me.

 

For example, I often take notes during meetings on my Moleskine and store it in Evernote. The Evernote OCR subsystem does a great job to make my notes searchable (even if I suspect that the text recognition is far from perfect, it is great for search). Often I then have to re-type parts of my notes to include them in documents, meeting notes, etc. If Evernote generated an automatic best-effort editable text that I can then copy and paste, it would go a long way to bridge the gap between paper and electronic notes storage.

 

Even though I understand the difficulty of making this into a feature and my expectations are low, I will still continue hoping that this feature will happen in the future It would change my way of using Evernote in combination with paper notes ;-)

 

Thank you.

 

No problem.  Evernote is becoming smarter and smarter every day.  Given time it may be able to run a meeting for you and take its own notes...  

[Frank.dg 1,385]

On 1/14/2014 at 2:31 PM, @lpasqualis said:

full results of the OCR analysis available as editable text

Looks like Evernote has its answer (for iPhone users) in the Evernote App store. 

I've tested out the iPhone app, eHighlighter - and I am thoroughly impressed! Based on comments in the (iOS) app store I forked out the $5 immediately.

Even though it says it only supports English (most likely the interface), I was able to take snapshots of portions of text from a variety of sources - in Portuguese - and have the parts I highlighted exported to Evernote.

So you go from this:

... to this:

 

In Evernote, you get the time stamp, page number (if you entered one) and a link to the original snapshot online.

In a matter of seconds you go from an image snapshot to editable text.

I was extra impressed that the Portuguese accents, inflections, etc. came through (as shown above). So English should be a breeze.

BONUS:

• Import images (with text) from your camera roll and convert those to editable text. So you're not limited to images shot within the eHighlighter app
• You could use your images saved to your camera roll (or Evernote) from Evernote's Scannable app...
• Save images already in Evernote to your camera roll and then access those through eHighlighter

 

 

[DTLow 5,736]

On March 8, 2016 at 7:44 AM, Frank.dg said:

I've tested out the iPhone app, eHighlighter - and I am thoroughly impressed! Based on comments in the (iOS) app store I forked out the $5 immediately.

Thanks for digging this up Frank - its a great tool (ehighlighter) and I paid for a copy.

[JMichaelTX 4,117]

On 3/8/2016 at 9:44 AM, Frank.dg said:

I've tested out the iPhone app, eHighlighter - and I am thoroughly impressed!

Does it support sharing, so that from the EN iOS app, I can tap an image, and then share the image with eHighlighter, which would then OCR the image?

[Frank.dg 1,385]

20 hours ago, JMichaelTX said:

Does it support sharing, so that from the EN iOS app, I can tap an image, and then share the image with eHighlighter, which would then OCR the image?

Nope. Not as far as I know. That would be nice, though.

It doesn't seem like one can access eHighlighter from the iOS share center (nor add it). The only way at the moment to get an image from Evernote to eHighlighter would be to (1) save the image to your camera roll, (2) open up eHighlighter and pull in the image from your camera roll. From time to time I see iOS apps rolling out the ability to share via the iOS share center, which is the ultimate. I hope it's in the pipeline... BUT, I must say that if one were predominantly to use the app to take snapshots of text from physical books, magazines, etc, it's pretty amazing:

1. Take a pic via eHighlighter
2. Mark the beginning and end of the phrase/ text you want to convert to editable text
3. Export to email, Dropbox or Evernote. 

Even doing things backwards to get something out of Evernote to the camera roll, etc. would be well worth it, considering the sheer amount of time saved in not having to transcribe your text manually. 

The other cool thing about this app is the ability to organize your snippets of converted text in categories... so you can read through a book and extract all the text you want, organized and sorted the way you want.

Except, @JMichaelTX, I'm not sure why one might want to use eHighlighter to take a snapshot and push it to Evernote for OCR when you could do that directly with the Evernote iOS app anyways.

[mistahchris 1]

I'd also love a feature to get editable text out of an image. Since there's no great way to do this right now, I've started using Google Drive's OCR capabilities then putting that text back into the note.

My process is:

• Use Evernote's Scannable App (which does an excellent job btw)
• It's set to always save to Evernote
• Send > Upload to Google Drive - this uploads the PDF or Image to a specific Google Drive folder. When I get time, I go into that folder, open the files with Google Docs and it converts it to editable text.

At some point I may investigate options with a Google App Script + EN API to get that text back into the right note automatically. I've looked into IFTTT and Zapier Options for this but they don't seem to offer the right combinations of triggers to get that process to work end to end.

 

Hopefully that Scannable process will be useful to some of you.

[Dave-in-Decatur 3,938]

I have an Android phone, and I use an app called Text Fairy: https://play.google.com/store/apps/details?id=com.renard.ocr. It can process an existing photo (if the photo is in Evernote, you can tap it and Save to make it accessible); or it can take a photo of a text document. It processes the text, and then offers various sharing options (through the Android Share feature), including creating an EN note. I've found this to work very well, for instance when doing research. The app is free and has no ads, and works in multiple languages.