作者: Yale Song
DOI:
关键词:
摘要: Esports has gained global popularity in recent years and several companies have started offering live streaming videos of esports games events. This creates opportunities to develop large scale video understanding systems for new product features services. We present a technique detecting highlights from game matches. Most use pronounced visual effects emphasize highlight moments; we CNNs learn convolution filters those highlights. propose cascaded prediction approach that allows us deal with challenges arise production environment. demonstrate our on dataset three popular titles, Heroes the Storm, League Legends, Dota 2. Our achieves 18 FPS single CPU an average precision up 83.18%. Part is currently deployed Yahoo Esports.