Script Based Pattern Recognition for a Portrait Tracking System on Broadcast Online Scenes

Abstract

The online broadcast scene is a novel scenario in culture and tourism industry in new period. However, the online broadcast control system has not been solved at the technical level. This research describes a collaborative smart control system for seizing actors and scenarios in broadcast online scene based on drama script with pattern recognition. Based on the script of the cultural and artistic forms under the online broadcast scenario, the training parameters of the pattern recognition algorithm are set to recognize the real-time capture of the face, dress, clothing, etc., and the tracking and overall setting control. The kernel algorithm is set with smart decision making system based for seizing actors and clustering algorithm (discrimination analysis) for seizing different scenarios on stage. Connections between equipment serving for the entire broadcasting drama and online presentations are achieved by the wireless network, software and kernel control system.