Paste Text from Image

[crutkas]

Sahil Malik has a wicked neat tweet. https://twitter.com/sahilmalik/status/1272988134570184704

leveraged this library. Totally doable.

• https://github.com/tesseract-ocr/tesseract
• https://winsmarts.com/clipboard-image-to-text-via-ocr-a65d4e7f6482

[barry-johnson]

This is a great idea! For me personally, I'd probably prefer to see it incorporated into the snipping tool.

[mahyar1284]

@barry-johnson yeah, it's a really great idea. like screen snip tool to select an area on the screen but it should recognize the text using a very simple image processing (OCR) and copies the recognized text to the clipboard.

[JoeyEamigh]

Yes - I would love to see this incorporated into the snipping tool - maybe some sort of keyboard shortcut to trigger the text copy instead of the normal screenshot. I would use this daily while taking notes on Teams lectures

[gileli121]

You don't need tesseract-ocr/tesseract
You should know that.. You are Microsoft.
You have OCR API built in the OS

See my answer here:
https://stackoverflow.com/questions/5079635/how-can-i-use-tesseract-ocror-any-other-free-ocr-in-small-c-project/58242313#58242313

Answer from gil123

[BillyFB]

There is already a good tool that does exactly this, have used it for quite sometime and it works really well. But would prefer something with better system integration, with less intrusive (cumbersome) usage, as now it works similar to tied to Snipping Tool as suggested, as it would always work also with Snipped content already if needed.

Text-Grab
https://github.com/TheJoeFin/Text-Grab

[barry-johnson]

@TheJoeFin Thanks for the note about the manner of activation/use of alternative OCR via changing the keyboard selection. For me, it would be great if I could select the OCR language within the tool, but switching keyboards pre- and post-capture would not be terribly burdensome, particularly as I recognize this would not be the standard use case for the vast majority of users. (FWIW, my real desire for use is to avoid needing to capture/download an image and send it into this image translator from Yandex: https://translate.yandex.com/ocr) which I would probably still use in some cases because it actually overlays the translated text in the image, so with things like charts and tables, the translations appear where they should.

And regarding the keyboard shortcut - I am fine with whatever the default is as long as it can be changed. For me personally, I would think about this as being closely linked to the screen snipping tool, so I would find the additional key toggle for invocation more logical than confusing.

And thank you for undertaking this!

[VasigaranAndAngel]

Also adding QR scanner and Visual Search with this will be very cool.

[Jay-o-Way]

We don’t have access to the snipping tool.

But we can access the Clipboard. (the image is sent to clipboard)

[Jay-o-Way]

@crutkas Shouldn't this title be "Copy text from image"?

[Jay-o-Way]

To add one more existing third-party suggestion: Project Naptha is a browser extention that puts ocr on images :)

[barry-johnson]

To add one more existing third-party suggestion: Project Nafta is a browser extention that puts ocr on images :)

Had not heard of this one. Thank you.

Project page: https://projectnaptha.com/

[Aaron-Junker]

@crutkas @TheJoeFin Many issues proposing an on-screen QR code feature are linking to this issue. Will this also be implemented in teh future?

[TheJoeFin]

@crutkas @TheJoeFin Many issues proposing an on-screen QR code feature are linking to this issue. Will this also be implemented in teh future?

I have not found a QR code reading .NET library which is offline and open source to use. If there is one then this feature could be easily added, otherwise maybe we should open a new issue for reading QR codes with PowerOCR.

[joaociocca]

(edit) Sorry, I meant to start by asking if I should actually post this here or open a new issue for it.(/edit)

I've noticed that when I select an image with text using different formatting (different colors, bold, etc), the tool change the order of stuff. For example:

I don't know what to write about, but I want an example. A really big and long example. I mean, not even the quick fox jumps over the lazy dog would be enough, probably.

if captured and pasted, would get pasted as...

I don't know, but I want an example. A really big and long example. I mean, not even the quick fox jumps over the lazy dog would be enough, probably.
what to write about

Also, it usually ignores yellow letters on a white background.

[Aaron-Junker]

(edit) Sorry, I meant to start by asking if I should actually post this here or open a new issue for it.(/edit)

I've noticed that when I select an image with text using different formatting (different colors, bold, etc), the tool change the order of stuff. For example:

I don't know what to write about, but I want an example. A really big and long example. I mean, not even the quick fox jumps over the lazy dog would be enough, probably.

if captured and pasted, would get pasted as...

I don't know, but I want an example. A really big and long example. I mean, not even the quick fox jumps over the lazy dog would be enough, probably.
what to write about

Also, it usually ignores yellow letters on a white background.

Hi @joaociocca

Please open a new issue

[TheJoeFin]

@Aaron-Junker Should this issue be closed since Text Extractor has been released?

[crutkas]

We need to close the 62 work items that are closed.

[crutkas]

@joaociocca please create a new issue with a sample image