VIDEO-BASED AUTOMATED GENERATION OF COMMENTARY IN CRICKET

Abstract

Throughout a cricket match, commentary entertains the viewers and gives updates about the game. Quoting pertinent stories associated to the current game situation makes the game more fascinating. The system proposes a solution which is an AI based approach that will assist the color commentators in effective storytelling that is interesting to the audience, and related to what is actually happening in the game being broadcast. The aim of the system is to retrieve and suggest stories that will assist the color commentator to describe appropriate remarkable incidents all through the course of the game. The system returns list of stories while a video is being played, based on how suitable they are for the game state. The methodologies used for system are video processing, video dataset, processing the dataset, Construction of a set of candidate stories, matching of video events and stories and finally generating commentary.