cancel
Showing results for 
Search instead for 
Did you mean: 

Google Vision Service API-Digital Exchange

ManojKumar3
Level 6
Hi Team, i am making use of Google Vision Service API from Digital exchange know as "Google: Computer Vision" where the action is  "document Text Extraction OCR". the issue here is i am getting output by using this service but then with 50% success rate which is very bad for this service . so want to make sure if anyone used this service and have good output or else if u can suggest some good service in Digital exchange .
5 REPLIES 5

ashish.easow
Staff
Staff
the success rate would probably depend on the documents which you have too, at most it would be suggestions but success rate you might have to determine by running the actual documents. Abbyy - PDF MS vision - image  Ephesoft - PDF

JamesW
Community Team
Community Team
Hi Some great info there. Would also suggest posting to the Blue Prism DX support forum for additional support. https://digitalexchange.blueprism.com/forums/forums/list.page  

Hi,
I am trying to configure Cloud Vision API in BluePrism. I have subscribed to Google Cloud Platform and created the service account. I configured the Credential with Private Key and email. But when I try to drag and use the Cloud vision API object, I could not see any action, from the below 8 services listed for this API. 
Detect Face
• Get Image Properties
• Safe Search Classification
• Label Entities
• Text Detection
• Logo Detection
• Landmark Detection
• Document Text Extraction (OCR)

I want to use OCR functionality. Can  you help me here. Did I miss any other configuration. Do we need to configure API separately.

------------------------------
Neha Tapkir
------------------------------

Hi Neha,

Which document are you referring to for the methods? For the Cloud Vision skill you're showing in your screen shot, each of those entries corresponds to an end points on the REST interface of that API. You can find more info in the Google docs here: https://cloud.google.com/vision/docs/reference/rest/

Cheers,

------------------------------
Eric Wilson
Director, Partner Integrations for Digital Exchange
Blue Prism
------------------------------

Neha,

Of the methods exposed by Cloud Vision, you want to focus on "vision.images.annotate". There's a single input parameter which is a JSON payload that you would build prior to invoking the request. The JSON payload will contain information that instructs Google as to what you're trying to do and the image to analyze. Here's an example:

{
  "image": {
    "content": "[Image to Analyse]"
  },
  "features": [
    {
      "type": "DOCUMENT_TEXT_DETECTION"
    }
  ]
}​


In this case "Image to Analyse" could be the value of an Image data item.

Cheers,

------------------------------
Eric Wilson
Director, Partner Integrations for Digital Exchange
Blue Prism
------------------------------