How to: Interpreting results from a Video Bounding Box Job

The output data of a video bounding box job is linked in the output column, the name of which is configurable.

The link will point you to a JSON file with results for the corresponding video. For every frame, there will be a description of every box in the frame (visible or hidden). You will also have an empty object for frames with nothing to annotate or with only hidden objects. Expect an array for every frame in the video. The schema is:

     "type": "box",
     "id": "605ed755-174b-4858-a3ae-dc241092c14d",
     "category": "Vehicle",
     "annotated_by": "machine",
     "x": 1156,
     "y": 257,
     "height": 14,
     "width": 14,
     "visibility": "hidden"
     "type": "box",
     "id": "ba92f510-4d56-417e-837d-d8814c8cabd7",
     "category": "Vehicle",
     "annotated_by": "human",
     "x": 675,
     "y": 307,
     "height": 14,
     "width": 8,
     "visibility": "visible"

The attributes are:

  • Type: This is the format of annotation, in this case, “box”
  • ID: The Unique and Persistent ID for the boxed object
  • Category: This is the class of the object from your ontology
  • Annotated_By : This field describes whether the box in this frame is an output of our machine learning (“machine”) or was adjusted / correct by a person (“human”)
  • X / Y: The top left corner of the box in question’s pixel coordinate X and Y
  • Height / Width:  The height and width of the box in pixels
  • Visibility: This will be “visible” when the object is in view and “hidden” when the object is hidden or out of frame.


Was this article helpful?
1 out of 2 found this helpful

Have more questions? Submit a request
Powered by Zendesk