Back to Search Start Over

VisDmk: visual analysis of massive emotional danmaku in online videos.

Authors :
Cao, Shuxian
Guo, Dongliang
Cao, Lina
Li, Shuo
Nie, Junlan
Singh, Amit Kumar
Lv, Haibin
Source :
Visual Computer; Dec2023, Vol. 39 Issue 12, p6553-6570, 18p
Publication Year :
2023

Abstract

Danmaku, a real-time comment function covering the top of the video, appears from the right of the video like a bullet and slides out horizontally from the left, which is gaining popularity in Asia. In recent years, the research on the analysis of massive danmaku data has mushroomed. The danmaku data contains a host of valuable information, such as the emotional expressions, attitudes, and opinions of the people watching the video, which helps people quickly get the content and effect of the video. The information is more representative and comprehensive with the ever-increasing amount of danmaku data over time. However, extracting valuable danmaku from huge amounts of data is a challenging task. Therefore, in this paper, we introduce VisDmk, an interactive visual analysis system, to help to analyze video content and effect. VisDmk incorporates five views: the projection view to exhibit emotion distribution, the detail view to analyze specific danmaku information, the individual view to illustrate the difference between viewers, the theme-aware view to identify themes in different periods, and the video view to ascertain some inference within a video. Case studies and user observation were conducted to evaluate this system. [ABSTRACT FROM AUTHOR]

Details

Language :
English
ISSN :
01782789
Volume :
39
Issue :
12
Database :
Complementary Index
Journal :
Visual Computer
Publication Type :
Academic Journal
Accession number :
173517645
Full Text :
https://doi.org/10.1007/s00371-022-02748-z