Semantic Analysis of Field Sports Video using a Petri-Net of Audio-Visual Concepts

摘要: The most common approach to automatic summarization and highlight detection in sports video is train an classifier detect semantic highlights based on occurrences of low-level features such as action replays, excited commentators or changes a scoreboard. We propose alternative the perception concepts (PCs) construction Petri-Nets, which can be used for both description event within videos. Low-level algorithms PCs using visual, aural motion characteristics are proposed, series Petri-Nets composed formally defined describe content. call this concept network–Petri-Net (PCN–PN) model. Using PCN–PNs, personalized high-level descriptions facilitated queries semantics achieved. A particular strength framework that we easily build detectors PCN–PNs search videos locate interesting events. Experimental results recorded data across three types games (soccer, basketball rugby), each from multiple broadcasters, illustrate potential framework.