Skip to content

Ouput Preview

Output Panel

The Output panel allows viewing the data that the solution is exposing. This tab can be accessed on the [Instance name] -> Output menu.

Output Panel

A new output group entry is created every time that the solution code writes to the Output Sink. For example:

instance:writeOutputSink("source", OutputType.Image, inputimg, "Input image")
instance:writeOutputSink("motion_pre", OutputType.Image, motionimg, "Motion debug (pre)")
instance:writeOutputSink("motion_post", OutputType.Image, motionimg2, "Motion debug (post)")
instance:writeOutputSink("detections", OutputType.BBox, detections, "Raw detections (untracked)")
instance:writeOutputSink("parking", OutputType.Poly, parking, "Parking spots")
instance:writeOutputSink("tripwires", OutputType.Line, tripwires, "Tripwires")
instance:writeOutputSink("tracks", OutputType.BBox, tracks, "Tracks")
instance:flushOutputSink()

Sink key

The first value passed to the writeOutputSink function is the key name of the sink. This key can be used in the output handlers to export data.

Output Types

Each output type creates a specific entry on the Output or Data panels. These can be:

  • image: Will create a new image source on the image dropdown list
  • bbox: Allows displaying and configuring bounding boxes on top the output image
  • poly: Allows displaying regions/zones on top the output image
  • line: Allows displaying lines/tripwires on top the output image
  • event: Allows displaying events in the Data panel
  • [custom]: You can also specify other keys that will not appear in the UI, but that can still be used to send data to Output Handlers and other modules

Data visualizer Tab

The Data visualizer Tab allows viewing all the Inference outputs and information on each entry. This tab can be accessed on the [Instance name] -> Data menu.

Events

Events

Clicking on the "Detail" icon, a popup window will show the JSON data of the event.

Mosaic

Mosaic

Table mode

cvediart-detector-inference.png

In the example above, the Data Visualizer is showing the result of the detection for a given frame.

Available information on Detector entries

  • ID : ID number associated with the entry
  • Label : Detector class associated with the entry
  • X/Y : Position values for the entry
  • Width/Height : Size values for the entry
  • Confidence : Confidence value associated with the entry
  • POT : POT value associated with the entry

Counting mode

Counting

Pixels On Target mode

POT