Pre-Trained Classifiers

OpenText may provide classifiers that you can use with Media Server to classify images.

The following classifiers are currently available, in the package MediaServerPretrainedModels_<VERSION>_COMMON.zip. When you download this package, ensure that <VERSION> matches the version of Media Server that you are using.

  • ImageClassifier_Broadcast.dat

    A classifier that can be used on broadcast footage to classify video frames into the classes "Adverts" and "News".

  • ImageClassifier_Document.dat

    Classifies images of documents into the following classes:

    • CV - a CV or resume.
    • Form - forms, questionnaires, and so on.
    • ID Document - passports, ID cards, driving licenses, security badges, and so on.
    • Leaflet - leaflets, flyers, brochures, information sheets.
    • Letter - typed letters.
    • Magazine - magazine covers and pages.
    • Newspaper - newspaper pages.
    • Poster
    • Report - such as technical reports.
    • Table - tabulated data.
  • ImageClassifier_Imagenet.dat

    Contains training for the Large Scale Visual Recognition Challenge (ILSVRC) classes listed at http://image-net.org/challenges/LSVRC/2012/browse-synsets.

  • ImageClassifier_RoadScene.dat

    A classifier for classifying road scenes. This classifier can classify images into the classes "car", "person", and "van".

  • ImageClassifier_TextFilter.dat

    Classifies images into the classes "Text" and "NoText", depending on whether they contain text. You can classify your images before running OCR, to avoid processing any images that do not contain text.

For information about how to import a classifier into your training database, see Import a Classifier