Automatic Video Summarization for Cricket Match Highlights using Convolutional Neural Network

Abstract

The technique of condensing a raw video into a more compact form without losing much information is known as video summarizing. This proposed system will summarize the sports match video to generate the highlights automatically. This tool will help the user to generate the video on giving an input of a sports match. Video analysis is the technique used by Video editors to edit, cut and merge the pieces of videos into fully summarized video this requires highly professional tools and professional knowledge of editing the video. Proposed system change the scenario of manual highlight generation to automatic highlight generation in some time and with more accuracy. Manual highlight generation necessitates expert editing abilities and is a time-consuming procedure, necessitating the development of a system capable of automatically generating sporting event highlights. Therefore, to distinguish and clip crucial events in a cricket match, it offer a system that takes into account both event-based and excitement-based attributes. Cues used to capture such events include audio cues, scoreboards, scene understanding, and player celebrations. The accuracy of proposed system to detect four is around 84.7%.