Text Semantics Based Automatic Summarization for Chinese Videos

摘要: This paper is concerned with the task of automatically summarizing videos, which important for many video-related applications. Examples include video retrieval, accessing and categorization, et al. Our approach summarizes semantically from texts embedded in videos. To improve accuracy, ant colony algorithm, maximum matching method, Singular value decomposition (SVD) are employed locating text region, segmenting words clustering sentences, respectively. Based on these methods, a prototype developed. In an experimental evaluation real world datasets, we show that our proposed could provide accurate, satisfactory performance.