Hi we try to read some text from screenshots,
but I recognize that the /vision/v3.2/ocr API has some issues recognizing text at all.
Is there an option to improve the API?
The following was not even in the result set: 

And here it cannot identify the rn and reads m

