The coordinates when using the cloud OCR appear to be after the correctSkew and any other pre-processing has happened. There appears to be no way convert these coordinates back to the coordinates of the source image... is that correct?
asked 23 Jun '13, 00:57
You are right, in Cloud OCR SDK the method that converts coordinates on a deskewed image to coordinates on an initial image is not implemented (it exists in FineReader Engine only).
If you image is rather straight, it is possible to disable automatic geometric correction: set - correctSkew=false, - imageSource=scanner.
This answer is marked "community wiki".
answered 24 Jun '13, 13:25
Anastasia Ga... ♦♦
has this behaviour changed? i need the opposite (actually get the corrected coordinates) and can't find a way or combination of settings to get an output with originalCoordinates="0"
answered 16 Sep '14, 14:16