While the use of natural language processing on “Bullet Screen” is beneficial to store and organize natural language, as authors provide users with search function to help users find the wonderful part of the video. According to Alexa, in January 2017, in all integrated authors site all over the world, Bilibili (http://search.bilibili.com/) ranked the 224th, while the number of its visitors ranked the 286th. It is the most active Chinese large-scale “Bullet Screen” video site. In this paper, research object is presented and related works about videos with “Bullet Screen” are detailed in 3. Research problems of the “Bullet Screen” retrieval system are discussed. Basic algorithms and the proposed ISB methods are demonstrated. Finally, the conclusions are provided. In fact, different types of “Bullet Screen” are applied in situations and have various characters. Subtitle is the result of the subtitle group (a small number of users) editing the video data. Most users use subtitle information rather than participate in the creation itself, which is similar to authorsb1.0. “Bullet Screen” is similar to Authorsb2.0, in which the users can be more interactive. In this process, the users can participate in the creation of the “Bullet Screen”. To some degree, texts “Bullet Screen” in live site have an impact on the broadcast itself. The anchor can communicate with the users through “Bullet Screen”. In this way, different “Bullet Screen” applications actually conform to the development of the Internet. Their development does not have a reciprocal relationship to each other. Whether it is subtitles, barrage or live barrage, it is not gone but applied in different situations. During the video viewing, the density of the “Bullet Screen” commentary is significantly correlated with the importance of the video
While the use of natural language processing on “Bullet Screen” is beneficial to store and organize natural language, as authors provide users with search function to help users find the wonderful part of the video. According to Alexa, in January 2017, in all integrated authors site all over the world, Bilibili (http://search.bilibili.com/) ranked the 224th, while the number of its visitors ranked the 286th. It is the most active Chinese large-scale “Bullet Screen” video site. In this paper, research object is presented and related works about videos with “Bullet Screen” are detailed in 3. Research problems of the “Bullet Screen” retrieval system are discussed. Basic algorithms and the proposed ISB methods are demonstrated. Finally, the conclusions are provided. In fact, different types of “Bullet Screen” are applied in situations and have various characters. Subtitle is the result of the subtitle group (a small number of users) editing the video data. Most users use subtitle information rather than participate in the creation itself, which is similar to authorsb1.0. “Bullet Screen” is similar to Authorsb2.0, in which the users can be more interactive. In this process, the users can participate in the creation of the “Bullet Screen”. To some degree, texts “Bullet Screen” in live site have an impact on the broadcast itself. The anchor can communicate with the users through “Bullet Screen”. In this way, different “Bullet Screen” applications actually conform to the development of the Internet. Their development does not have a reciprocal relationship to each other. Whether it is subtitles, barrage or live barrage, it is not gone but applied in different situations. During the video viewing, the density of the “Bullet Screen” commentary is significantly correlated with the importance of the video