Paper Title
Score Point Event Extraction From Tennis Video Using Embedded Text Detection

Abstract
In a sports video there are many events to be found to make highlight of a game. The event is a one which has important scenes which are very interesting to watch. Manual work of cutting important events and editing is very common and not much technology or concepts are required. For automatic identification of events in sport video need some technique to solve this problem. A novel system for tennis sport video event detection using embedded text detection for construction of sports video summary is proposed. The system has two components: event detection and separating frames around detected event. Event detection is realized by change detection of game stat texts (such as text information showing in score box). The developed system detects the game points as the event beginning and potential endings of the event. Then frames are separated around detected event.