Here’s a reminder that as a registered user of Virginia Chronicle, the Library of Virginia’s digital newspaper collection, you can assist in improving search results by correcting inaccurately translated newspaper text.
Optical Character Recognition (OCR), the software which reads a scanned newspaper page to create a searchable text file, is not 100 percent accurate. Many things can affect OCR accuracy, including broken or blurry type, text that is too dark or too light, mixed fonts, etc. Therefore, we need users to correct words the human eye can read that OCR cannot.
To become a registered user, go to the Virginia Chronicle page and click “Register” in the upper right corner of the home screen and enter the necessary information. Once you are registered, you will need to log in with your email and password whenever you would like to correct text in Virginia Chronicle.
Next, go to the newspaper page you would like to correct, right click on the page and choose “Correct page text” from the three options. There is also a “correct this text” option on the left side of the screen under “Why may this text contain mistakes?”
A “Correct Text” column appears on the left with text that has been read by Optical Character Recognition software while the newspaper image displays on the right. You can place your cursor anywhere under the “Correct text” column on the left and begin text corrections–a corresponding red box will appear over the newspaper image to show you where you are on the page. Before leaving each section–usually pages are sectioned by column indicated by a blue block–click the “save” button to save any changes you’ve made.
And that’s all there is to it! You’ll see that the it’s a fairly easy process once you actually begin correcting text.
You can also click on the “How to correct text” link or to the “Help” feature for detailed instructions as well.
The Virginia Newspaper Project would like to extend a big thank you to the registered users of Virginia Chronicle who have already corrected over 300,000 lines of text!
Thanks to engaged users like them, this great research tool is made even better. Thank you!