3

I'm trying to recognize vertical text using google cloud vision. Image example: A B C D E

I use Try This API on https://cloud.google.com/vision/docs/ocr to test the engine.

Request body:

{
  "requests": [
    {
      "features": [
        {
          "type": "DOCUMENT_TEXT_DETECTION"
        }
      ],
      "image": {
        "source": {
          "imageUri": "https://i.stack.imgur.com/3wwYp.jpg"
        }
      }
    }
  ]
}

The result is

{
  "responses": [
    {}
  ]
}

Am I missing something? Thank you.

ThomasEdwin
  • 2,035
  • 1
  • 24
  • 36

2 Answers2

1

You can try to set "model": "builtin/latest" as per the documentation, that will give you some results. I believe recognising vertical text is something that is being worked on now, so you should not expect 100% accuracy at this moment.

Another issue is that single character detection is currently not supported by the Vision API.There is a feature request submitted with regard to character recognition here. Please star it so that you could receive updates about this feature request and do not hesitate to add additional comments to provide details of the desired implementation.

You can also have a look at this post, where people discuss vertical text recognition.

I have tried my first suggestion with your image, and for some reason A is not recognised, perhaps the Vision API sees it as "being separate" from the rest of the letters, which form a sequence, and is not being recognised since it is treated as a single character in that case.

Philipp Sh
  • 967
  • 5
  • 11
0

You may want to try another feature. For example

"type": "TEXT_DETECTION"
Wakanda
  • 26
  • 5