News Video Indexing and Story Unit Segmentation using Text Cue
international conference on wireless communications and signal processing(2019)
摘要
News video Indexing and story segmentation are one of the important steps for the creation of well categorized, organized and indexed video storage system. Such a system allows disciplined storage to cope with an increasing amount of data, and easier retrieval in the future for viewing. Keeping this approach in mind we proposed a News video Indexing and Story unit segmentation algorithm based on keywords obtained from the text displayed in the news frame during a news program. First with the help of our shot categorization algorithm we classify the news shots into commercial and news based on detection of spatial features of the particular News Channel. Then with the help of our shot indexing algorithm, we detect the presence of words in each frame of a shot and index each shot with a keyword. With the help of our Shot Combination algorithm, we combine the shots having similar keywords to form individual stories. The results obtained in this works were satisfactory and in accordance with the results depicted using manual visual observation.
更多查看译文
关键词
Video Indexing,Story Segmentation
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络