How the audio/video content is automatically taxonomized automatically?

Wednesday, 02 February 2011 13:07

Semaphore classification is based on a Natural Language Processing engine. We can extract any text metadata associated with image and video formats that support some form of text property, or often the path and filename can give us an indication. This can be assessed by the rules based engine. A different classification strategy can be employed for images to compensate for the relatively small amounts of text evidence available.