Overview

This node tracks the specified attributes for objects in the video. It can keep, remove or persist attributes from objects after they are seen for a certain duration. Use it to smoothen out object attributes when detections are intermittent.

Requires a AI Model Node before it in order to function properly. Having a Track Objects Node before is required if using Object Match Patterns else you can place it after the Merge Objects node.

Inputs & Outputs

Inputs : 1, Media Format : Raw Video
Outputs : 1, Media Format: Raw Video
Output Metadata : Filtered Objects Information

Properties

Property	Description	Type	Default	Required
`object_types`	If specified, only process objects of these types. Leave blank to process all objects.	model-labels	`null`	No
`object_match_patterns`	Only modify objects with IDs specified in the metadata attribute(s) defined by this pattern. Requires tracker prior to this node. Format: comma-separated metadata paths or patterns.	`string`	`null`	No
`tracked_attributes`	Format: `<attribute>` (specific attribute) or `*` (all attributes). Ex. `wrong_direction,object_of_interest`. Keep/remove the specified attributes from object after they are seen/not seen for a certain duration.	`string`	`null`	Yes
`min_presence_secs`	Only process objects/attributes that have been seen for at least this duration. Unit: seconds.	`float`	`0`	No
`attribute_buffer_secs`	Keep the Tracked attributes on a object until they are not seen for this duration. Unit: seconds.	`float`	`0`	No
`object_history_duration`	Remember objects for this duration after they are last seen in case they reappear. Unit: seconds.	`float`	`600`	No

Output Metadata

The fields below are declared by this node's metadata schema; the JSON values are representative examples.

Path	Type	Description
`objects`	`array`	Value of objects.
`objects[].attributes`	`array`	Value of attributes.
`objects[].attributes[].class_id`	`integer`	Value of class id.
`objects[].attributes[].label`	`string`	Value of label.
`objects[].attributes[].probability`	`number`	Value of probability.
`objects[].class_id`	`integer`	Value of class id.
`objects[].id`	`integer`	Value of id.
`objects[].label`	`string`	Value of label.
`objects[].probability`	`number`	Value of probability.
`objects[].rect.height`	`integer`	Value of height.
`objects[].rect.left`	`integer`	Value of left.
`objects[].rect.top`	`integer`	Value of top.
`objects[].rect.width`	`integer`	Value of width.

JSON example

{
  "objects": [
    {
      "attributes": [
        {
          "class_id": 100,
          "label": "hat",
          "probability": 0.8
        }
      ],
      "class_id": 1,
      "id": 101,
      "label": "person",
      "probability": 0.9,
      "rect": {
        "height": 100,
        "left": 20,
        "top": 20,
        "width": 50
      }
    },
    {
      "attributes": [
        {
          "class_id": 100,
          "label": "hat",
          "probability": 0.8
        }
      ],
      "class_id": 1,
      "id": 999,
      "label": "person",
      "probability": 0.9,
      "rect": {
        "height": 100,
        "left": 20,
        "top": 20,
        "width": 50
      }
    },
    {
      "attributes": [
        {
          "class_id": 100,
          "label": "vest",
          "probability": 0.8
        }
      ],
      "class_id": 1,
      "id": 404,
      "label": "person",
      "probability": 0.9,
      "rect": {
        "height": 100,
        "left": 20,
        "top": 20,
        "width": 50
      }
    }
  ]
}

Additional details

Object labels and attributes

Object labels/classes added: None.
Object attribute labels/classes added: Previously observed upstream attribute labels selected by attributes_to_track; restored entries use class 45000.