Object Recognition

Object Recognition identifies objects in video and image files and displays when within the media each object is identified. It is an API-enabled service processed by Google Cloud Vision, which has more than 20,000 objects in its database and is updated regularly.

Object recognition provides an additional search mechanism to spoken word text transcripts. The main use case for object recognition is to look at context of where the feedback is being given. Video is not currently able to recognize brand names or logos.

Restriction: Video Analytics are not available for Video Experience Edition. Contact your Medallia Representative for more information.
Note: Object Recognition can be activated for Video Research Edition and Insights Suite customers. Contact your Medallia Representative for pricing and activation information.

Visualization

Objects identified in video are displayed in the Object timeline chart on the player page in Video. By default, Video displays the five objects that appear most frequently throughout the video.

A screenshot of the LivingLens Object Recognition chart. It displays five objects on a clickable timeline.

Additional objects are nested beneath the top five.

To access additional objects:

  1. Click Down arrow menu icon Expand Objects.

  2. Click an object to view it on the timeline.

  3. The selected object appears at the top of the Object timeline. Click anywhere on the Object timeline to jump to the moment in the video.

Objects identified in Image files are displayed beneath the image in Video.

The Analytics page can display an aggregate view of up to 70 most frequently identified objects in a word cloud. The word cloud display is weighted by the number of media items the object appears in. For more information, see Analytics Module display.