Calculate Scene Metrics

Pre-requisites

You must have an existing video with an associated media_id and the prediction status as done.

Use the media_id to send the POST request to the endpoint /predict/v2/videos/media_id/scene_detection

Metrics can be calculated for predefined scenes or, in case no scene transition frames are provided, the scene detection is automated,and scores are presented from the defined or auto-detected scenes.

Expected request payload for predefined scene metrics

The scene_transition_frames are optional.

{
    "scene_transition_frames": [ 353, 578 ]
}

The transition frames signify the "boundaries" of each scene. So if, for example, in a video with 800 frames, the provided transition frames are 353 and 578, as showcased above, then the metrics will be calculated for 3 scenes:

From frame 1 until frame 352
From frame 353 until frame 577
From frame 578 until frame 800

For automated scene detection, no payload should be sent.

Example of a successful response

The results present the average of the prediction metrics for each scene. Each scene is defined by a unique identifier in the form of a UUID.

{
  "media_id": "c88ded20-69ad-457e-ae34-d04dfa9ad447", 
  "success": true, 
  "message": "OK", 
  "results": {
    "684b309b-5c75-434f-8db7-3f94f1bdda75": {
      "start_frame": 1, 
      "end_frame": 352, 
      "focus": 80.50604034612387, 
      "cognitive_demand": 45.078818846324495, 
      "memory": 56.21344592314347, 
      "engagement_frt": 62.88968305369173, 
      "aois": [
        {
          "uuid": "b8eded7a-1d2d-5452-df39-4fa7298ab972",
        	"name": "Getty Images",
       		"availability": 0.08333333333333333,
        	"total_attention": 23.5226571559906,
        	"relative_attention": 23.5226571559906          
        }
      ]
    }, 
    "08244ec2-0e41-40ea-8341-364415c452a9": {
      "start_frame": 353, 
      "end_frame": 577, 
      "focus": 87.38416864131615, 
      "cognitive_demand": 47.39792340437856, 
      "memory": 58.96938118994907, 
      "engagement_frt": 59.803821352594724, 
      "aois": [
        {
          "uuid": "b8eded7a-1d2d-5452-df39-4fa7298ab972",
          "name": "Getty Images",
          "availability": 0.7083333333333334,
          "total_attention": 31.4802991993287,
          "relative_attention": 31.4802991993287          
        }
      ]
    }, 
    "f7b93f99-cd0a-4627-986e-a7b46644c3df": {
      "start_frame": 578, 
      "end_frame": 800, 
      "focus": 68.33355230263865, 
      "cognitive_demand": 72.30813829044762, 
      "memory": 72.1463888155716, 
      "engagement_frt": 49.55575712129309, 
      "aois": [
        {
          "uuid": "b8eded7a-1d2d-5452-df39-4fa7298ab972",
          "name": "Getty Images",
          "availability": 0.5833333333333334,
          "total_attention": 0.6565483934853388,
          "relative_attention": 12.94338261442525
        }
      ]
    }
  }
}

In videos containing AOIs (static or autodetected) the scene metrics also contain the availability, total and relative attention of each video AOI per scene.