Please use this identifier to cite or link to this item: https://dl.ucsc.cmb.ac.lk/jspui/handle/123456789/4640
Title: Automated Highlights Generator for DOTA 2 Game Using Audio-Visual Framework
Authors: Sanjeewa, K.W.C.K.
Issue Date: 26-Aug-2022
Abstract: This study shows the automated highlight generation through the audiovisual framework for esport Dota 2. Currently, in most of the video sharing platforms there are plenty of content creators on top of e-sports. Those contents are mostly based on the games played by the players. Community of the e-sport tent to watch those content through those platforms. With the high availability of contents and lack of time community missed most of the interesting videos. While discuss with the community can see that they tend to watch the highlights which is shorter version of full game play. While surfing video sharing platforms such as YouTube, stats shows that there are more views for highlights videos than full videos. Generation of highlights of the video consists of creating a shorter version of the video which consists of the most interesting parts of the original. It requires domain knowledge and timeconsuming process. In order to reduce those effort this research will facilitate efficient highlight generation tool with highlight detection and give final output as MP4 video output. This research work mainly based on the audio intensity detection, template matching to detect key events and event detection using predefined text appearing on the screen. In this research jointly uses visual features and audio features to construct highlight generation model and enable compact highlight representation. In this highlight generation process, first identified the clips by analyzing the energy of the audio file. Identified clips passed into visual processing module. In the visual processing module that uses template matching and OCR techniques to identify the highlighted clips. Finally, all identified clips have been merged and created output in video file format.
URI: https://dl.ucsc.cmb.ac.lk/jspui/handle/123456789/4640
Appears in Collections:2021

Files in This Item:
File Description SizeFormat 
2018 MCS 081.pdf1.3 MBAdobe PDFView/Open


Items in UCSC Digital Library are protected by copyright, with all rights reserved, unless otherwise indicated.