Object Detection (Amazon Rekognition)


Submits video asset files to Amazon Rekognition service to detect on-screen objects. Results are added as temporal markers to the asset version.


Can be used to detect the following on-screen elements:

  • Celebrities
  • Content Moderation
  • Faces
  • Objects
  • Shot Segments
  • Text

Requires a H.264 browse file. The default minimum confidence is set to 90.0%. The allowable range is 50.0% – 100.0%