Recognise text from frames or videos

  • Last Post 23 November 2017
  • Topic Is Solved
Alejandra Garrido posted this 05 October 2017

Good morning,

We are doing joint tests with different OCR systems and we need to process the same frames. So, I was wondering how it could pass directly to the Real Time Recognition path where to extract the frames to be processed.

Order By: Standard | Newest | Votes
Diana Khammatova posted this 11 October 2017


RTR enables your application to capture information directly from the preview frames of your smartphone camera, without actually snapping a picture. Briefly, you need to implement the Callback interface which will be used to pass the data to and from the recognition service. Here are the recommendations on what the interface's methods should do:

onRequestLatestFrame method should retrieve the image from the camera and pass it on to the IDataCaptureService.submitRequestedFrame method.

onFrameProcessed method is where you work with the results.

Full scenario on how to capture information directly from the frames please find in the Help → Guided Tour → How to Capture Text from Camera


  • Liked by
  • Alejandra Garrido
Diana Khammatova posted this 23 November 2017

We would like to add that it is possible to save frames in the raw NV21 format, then fill the buffer with the saved data (and not pass it to the camera) and use the saved buffers or buffers converted to BMP or PNG with other products.