Geek Answers Handbook

Can I get boudingPoly for LABEL_DETECTION results?

How can this be done using the Google Vision-API, please?

  • send image to vision-api
  • request: 'features': [{': 'LABEL_DETECTION', 'maxResults': 10,}]
  • receive shortcuts, in particular those in which I am interested, is a "watch"
  • get boundingPoly so that I know the exact location of the clock in the image
  • having received boundingPoly, I would like to use it to create a dynamic AR marker that will be tracked by the AR library

Currently, the Google Vision-API does not seem to support boudingPoly for LABELS, so the question is whether there is a way to solve it using the Vision-API.

+4
source share
2 answers

Label detection does not currently provide this feature. We are always considering ways to improve the API.

+2
source

Two years later, the same thing. I am facing similar problems and I am thinking about choosing other solutions. I think custom solutions like TensorFlow object detection APIor DarkNet YOLO object APIwill do the job very easily.

0
source

More articles:


All Articles