Core Concepts
Short videos' affective computing is crucial for public opinion guidance and MSEVA offers a solution for emotion analysis.
Abstract
The article introduces MSEVA, a system for analyzing emotions in short videos. It highlights the rising demand for short video content and the impact of emotions on public mood. The system includes a multimodal dataset creation, automatic audio segmentation, and an improved affective computing model. MSEVA aims to provide timely public opinion guidance by analyzing emotions in short videos.
Structure:
Introduction to YouTube Shorts as a competitor to TikTok.
Importance of emotional analysis in short videos.
Creation of a multimodal dataset for emotion analysis.
Proposal of the MSEVA system for emotion analysis in short videos.
Key contributions and results of the study.
Stats
この論文では、感情分析モデルの精度を約4.17%向上させることが示されています。
この論文は、147本の短いビデオからなるデータセットを構築しています。