Document Understanding- OCR Engine Options

Help Activities
, ,
1

Hi Team,

For one of the use case we are using UiPath Document Understanding Framework with UiPath OCR Engine/ Omni page OCR engine in Digitization step, But somehow as the images are distorted so the Digitized Text is not upto the mark. As there any other better OCR engine we could use or any Other technology instead of UiPath DU?

Thanks in Advance !!

2

Hi,

We can also use GoogleCloudVision or Azure Computer Vision etc as the following.

image
image1112×405 30.9 KB

Regards,

1 Like
3

@Yoichi : Sure, but do you have any idea which one is best for distorted images as per your experience? as UiPath OCR & Omnipage result is not upto the mark.

4

Hi @avinash_ghanwat1

The best is to try different OCR engines on a selected set of sample documents to see which one provides the best output for your documents. Since the document quality is not that great, testing a few and analyzing the results is always good.

I would also suggest looking at methods to improve the document quality. Maybe grayscaling or some enhancement steps. If possible, also consider adding some standards to the process that describes how they should prepare the documents for the automation.

1 Like
5

Hi @avinash_ghanwat1

I have worked on similar useacse where I have tried multiple OCR’s including microsoft and google .But among these all, UiPath DU OCR was more efficient.

1 Like