ORC Project help

  • Last Post 14 April 2016
johnathan posted this 14 April 2016

We're having an OCR project and need help with recognizing the text/serial number from the attached images. Please let me know if you guys can do it and feel free to contact me if you need further information photo1 photo2

Oksana Serdyuk posted this 14 April 2016

ABBYY Cloud OCR SDK can be used for your usage scenario, but at first it is necessary to improve the source images quality and try to find more optimal recognition settings for your kind of images. Below you can find some our tips for taking photo:

  • Make sure that the page fits entirely within the frame on the screen while taking a picture. It is very important the image contain only the text the customer wants to recognize, other objects and colored background can lower the recognition quality.
  • Make sure there is enough light (preferably daylight). In artificial lighting, use two light sources positioned so as to avoid shadows.
  • Turn off the flash to avoid glare.
  • The text on the image has to be in focus. Enable the anti-shake system, if available.
  • The text on the resulting photo should not be skewed by more than 20 degrees.

Also please refer to the Source Image Recommendations section where you can find more recommendations.

Please take a look at the processTextField method and the article with some details in documentation. Also it can be also very useful to limit characters with the help of the letterSet parameter in your case, for example: Letterset = "-.=JGRPreldB0123456789". You can see this method implementation and test its work using our C# GUI Sample > select the “Field-level” tab.

I have tested your images and managed to extract some text from the first photo:

alt text

Attached Files