Train Media Server to Detect Objects

Before using object detection, you must import a detector that is provided by HPE. The detectors provided by HPE are pre-trained, so you do not need to perform any training. For information about the detectors that are available, see Pre-Trained Object Detectors.

The following diagram shows how Media Server stores information for object detection:

The "Media Server database" represents the Media Server datastore file or a database on an external database server. For information on setting up this database, see Introduction.

The Media Server database can contain any number of detectors. An object detection task uses a single detector and finds instances of objects that are described by classes in that detector. For example, if you are processing video captured by a CCTV camera, you could use the roadscene detector which contains classes for finding people, cars, and vans.

A detector contains a background class and a minimum of one foreground class. The background class contains images of uninteresting (background) objects and this class is never returned in a detection result. Only foreground classes are returned in detection results as they represent the categories of objects that you want to detect.