首页|基于弹幕数量变化预测视频切片精彩程度的研究

基于弹幕数量变化预测视频切片精彩程度的研究

扫码查看
随着视频行业的飞速发展,关于视频精彩程度的分析方法也越来越多.使用计算机视觉和音频处理技术的方法消耗成本太高,而基于用户观看行为的评估方法则对视频具体片段的精彩程度评估有所忽略.弹幕作为一种新的视频评论形式,在以视频具体片段为反馈对象的同时,也为原本的视频内容提供了额外的信息.收集超过 8000 条包含 2 分钟的视频切片及对应弹幕的数据集,并基于此数据集提出了一种用于评估视频切片精彩程度的新方法.经过评估,此方法有效.
Research on Predicting the Interestingness of Video Clips Based on the Change of the Number of Barrages
With the rapid development of the video industry, there are more and more methods for analyzing the interestingness of videos. As the cost of using computer vision and audio processing techniques is too high, and the evaluation methods based on user viewing behavior neglects the evaluation of the interestingness of video clips, barragesas a new form of video commentary have emerged.Barrages not only provide feedback on specific video clips, but also provide additional information to the original video content. In this study, more than 8,000 pieces of data that contains two-minute video clips and corresponding barrages were collected. A new method has been proposed to evaluate the interestingness of video clips. This method has been evaluated as effective.

BarragesVideo clipsVideo interestingnessClassification models

余彦子、唐杨

展开 >

五邑大学 文学院,广东 江门 529000

航天中认软件测评科技(北京)有限责任公司 北京 100000

弹幕数据 视频切片 视频精彩程度 分类模型

2024

五邑大学学报(社会科学版)
五邑大学

五邑大学学报(社会科学版)

CHSSCD
影响因子:0.146
ISSN:1009-1513
年,卷(期):2024.26(2)
  • 8