How To Extract Text From Images (OCR) – Windows Only
OCR (Optical Character Recognition) is cool – the thing that lets you analyze an image and recognize the text within the image. Something that you can do with Evernote. However what if you want to do it for some of your own images? Of course you can use Evernote or you can use JOCR.
A pretty straight forward and no fuss application. It does one thing and does it well. Download the application and install. Fire it up and you will be greeted by the main interface.
JOCR essentially allows you to take screenshots and then detect the text within the image, excellent for situations where you are copying from an image but you don’t want to type the text.
You can capture a region, a window or the entire desktop and then perform character recognition. Once you have captured the image just hit recognize and JOCR will attempt to give you the text from within the image.
It is pretty fast and the actual recognition time will depend upon the type of image you are trying to perform OCR on. The recognized text is placed on the clipboard and JOCR will open up a notepad file where you can paste the text and see how well JOCR fared. As you can see it did fairly well (just messing up Windows to IWindows in line 3)
Although JOCR is officially limited to reading screen grabs, however there is nothing stopping you from scanning a document, save it, open it up at 100 % and summon JOCR. It actually did a pretty good job when I opened this image and then took a screenshot of the window to read it.
JOCR requires that you have Microsoft Document Imaging installed on your system. If you have Office 2003 or higher, chances are that you have it installed. In case it is not, you can do so from within “Add and Remove Programs” and then choosing to install “Document Imaging”.
JOCR can recognize text in a number of languages ranging from Dutch to Hungarian to German to Russian. You would have to choose the language manually depending upon the type of characters you are going to be recognizing.
Although the application has not seen updates since some 2 years, it does a pretty good job on both black and white as well as colored images and could be a time saver with your scanned documents or images.
Have you ever used an OCR application? Which ones do you know about?
(By) Varun Kashyap - Programmer, Blogger and Tech Enthusiast, who tweets @VarunKashyap and blogs about tips, tricks and latest on the web at TechCrazy Blog
Nice software. I will test this
(Comments wont nest below this level)Till now I tried OCR that came with Scanner and tried Simple OCR,I will have to try JOCR at some time .
(Comments wont nest below this level)Does not work. But it would be very cool!
I get this error message :
The error message is ActiveX component can’t create object
Is it proper string image?
Or, confirm your available language by executing
[MS Office > MS Office Tools > MS Office Document Imaging]
and selecting [Tools > Options > OCR tab > OCR Language]
MS Office Document Imaging is installed with my office XP.
Document imaging version is 1.03.2349.1 what is your office version ?
Thanks
(Comments wont nest below this level)Make sure that you have Microsoft Document Imaging installed on the system. And if you have make sure that the language you are trying to read is available with Microsoft Document Imaging.
eg if you want recognize German language text, make sure you have support for it
i get the same error as above…
(Comments wont nest below this level)Yes it is ! English is selected in OCR settings in Document imaging and I’m trying to OCR some english text.. too bad it is not working.. I did not try Evernote but it does not claim to be doing that..
(Comments wont nest below this level)Tools is excelent thanks .
(Comments wont nest below this level)Do you know of any OCR tool that you can actually train to recognize certain image to a specified character? It would be powerful if you could create a custom OCR engine, for any language.
(Comments wont nest below this level)I too got the same error
(Comments wont nest below this level)