Discovering attractive segments in the user-generated video streams

作者：

Highlights：

• We build a dataset, which includes video and time-sync comments, for segment popularity prediction.

• We study the problem of video segment popularity prediction, which can help audiences to find attractive shots.

• With the help of time-sync comments information mining, the model can better understanding the audiences.

• Combining visual and text information can make results more precise.

摘要

•We build a dataset, which includes video and time-sync comments, for segment popularity prediction.•We study the problem of video segment popularity prediction, which can help audiences to find attractive shots.•With the help of time-sync comments information mining, the model can better understanding the audiences.•Combining visual and text information can make results more precise.

论文关键词：Time-sync comment,User-generated video stream,Segment popularity prediction,Video-to-text transfer

论文评审过程：Received 14 July 2019, Revised 4 September 2019, Accepted 16 September 2019, Available online 25 September 2019, Version of Record 25 September 2019.

论文官网地址：https://doi.org/10.1016/j.ipm.2019.102130