Train Media Server to Classify Images

To classify images, you must train Media Server by providing images that are representative of your chosen classes.

The following diagram shows how Media Server stores information you provide during training.

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

A classifier contains the object classes that you want to use for a classification task. When you run image classification, Media Server classifies your images into the classes in your chosen classifier. For example, to classify vehicles into "cars", "trucks", "motorcycles", and so on, you would create a classifier named "vehicles" and create object classes named "car", "truck", and "motorcycle".

A classifier must contain at least two object classes. To determine whether an image contains a car or doesn't contain a car, you would need to train a "car" object class and also a "not car" object class.

To each object class you must add training images. Usually around 100 training images per object class is sufficient to obtain good performance. For information about choosing suitable training images, see Training Requirements.